The venerable mainframe is so deeply embedded in many businesses that it’s hard to imagine eliminating it. Countless person-hours and custom code have been invested — for over half a century, in some cases.
Nevertheless, myriad factors, including the shift to a digital-first attitude in business, are prompting many businesses to reconsider their technologies and processes, including how to address aging legacy platforms and applications. We are frequently asked by clients about various options for the workhorse mainframe and how they might modernize its inner workings. Businesses wonder if they should update the platform itself; adapt new approaches for the data and applications that run on it; or move on altogether.
This three-part series applies our experience and shares client examples to answer these questions. We begin by explaining the factors that enterprises should use in deciding whether the mainframe still suits their needs. Part 2 will inform companies about retaining and optimizing their mainframes by exploring how to make this happen. Finally, part 3 looks at the future of the mainframe in an increasingly digital business world.
Sometimes, it seems “the mainframe is dead” predictions go back almost as far as the platform itself. But these predictions are as wrongheaded as they are cliched. According to a 2019 report from Allied Market Research, the global mainframe market was just under $2.1 billion in 2017 and is projected to hit nearly $3 billion in 2025. That’s a projected compound annual growth rate of 4.3% from 2018 to 2025 — a mature market, to be sure, but not deceased by any means.
And that’s just part of the picture. While the quantity of mainframe use is important — especially in this age of big data, which isn’t going away in the foreseeable future — the nature of mainframe reliance is vital as well. The platform has long been relied on to manage the operational data that forms the lifeblood of business in secure, reliable, economical fashion. As noted, heavy-usage organizations have spent decades and millions customizing their environment and apps to serve company-specific purposes.
The great rethink
Despite the many strengths of mainframe computing, the platform is under scrutiny for several reasons, most of which align with one of two budget categories: cost and digitization.
- Cost factors. This not only includes the basic cost per million instructions per second (MIPS), but also the move to cloud (often driven by a desire to replace the mainframe’s significant capital expenses with an as-a-service model); data center footprint and capacity; and total cost of ownership.
- Digitization factors. These include the need for greater agility; the drive to automate processes and the cloud-native development model involving continuous delivery, DevOps and microservices; the growing imperative to study and engage with customers 24x7 across multiple channels; and the COVID-19-fueled growth of digital initiatives and hybrid work environments.
With these factors in mind, enterprises are taking a hard look at their mainframe services’ limitations and weaknesses. Mainframe concerns vary depending on the size of the organization (see below). Smaller enterprises — which, for the purposes of this discussion, we define as those using fewer than 120 million service units (MSU), which equates approximately to 1,000 MIPS — are paying very high prices for that computing power. This leads to a powerful fiscal argument against the platform, so it’s little wonder that these businesses’ most common question is how to exit the mainframe.
Note, however, that even for smaller users of mainframes, the decision is seldom a simple dollars-and-cents issue; organizations of all sizes should go through an assessment (see part 2 of this series) of mainframe benefits, pain points, potential reasons to exit or stay in the platform, and whether to modernize it to create a “digital mainframe” platform, which is characterized by one or more of the following:
- Adoption of traditional and new-age applications running on the mainframe
- Adoption of DevOps
- API-based integration of legacy applications and data to channel applications
The figure below illustrates generally accepted industry figures on how many MIPS are considered small, medium or large mainframe usage.