Greening the Mainframe Grass: How to Optimize ROI for Mainframes

If you drive around suburbia in the summer, you’ll probably notice that some lawns are greener than others.

It’s not because the grass on those lawns is genetically superior, or because the soil on one side of the fence is richer than the other. Nor does one lawn get more rain or sunlight than the lawn next door.

Instead, the difference between a lush, green lawn and a dusty, scrabbly lawn hinges on how well each lawn’s maintainer invests in optimizing what the lawn produces. It’s all the same grass; what matters is how homeowners take care of their grass.

What color is your mainframe?

Mainframes are similar to lawns in the respect that it’s not the underlying mainframe infrastructure that determines how well your mainframe functions. It’s the way you use that infrastructure, and how much care you put into optimizing it.

After all, most mainframe systems today look more or less the same from a hardware perspective. They’re also similar from a high-level software perspective, because almost all of them run the same IBM operating systems.

Yet, if you look around the mainframe landscape today – as we often do at DataKinetics as part of our mission to help businesses optimize mainframe performance and cost – you’ll notice that there is a huge variation between companies with regard to how much value they get out of their mainframes. Some mainframes are the equivalent of beautiful, well-maintained lawns, while others are parched patches struggling to justify their purpose.

Understanding the reasons behind these different outcomes is critical for any business that wants to make the most of its mainframe systems. And given that mainframes are used by 71 percent of global Fortune 500 companies – and that they power 68 percent of the world’s transactions – learning how to optimize mainframe performance while minimizing cost is a clear priority for businesses across the planet.

To that end, this article unpacks the main factors that shape mainframe outcomes. In other words, it explains how to optimize the ROI of mainframes by deploying cost-management and optimization strategies that maximize the speed and performance of mainframe workloads while minimizing energy, maintenance and integration costs.

Mainframes want to save you money…

The most important thing to know about mainframe cost optimization is that mainframes are designed to be cost-effective. Consider data points like the following:

• Mainframes consume 50 percent less energy than x86 machines when running workloads of equivalent capacity.
• Mainframes take up less physical space than commodity servers, which leads to lower real estate costs.
• On the whole, labor costs for mainframe maintenance are 50 percent lower than for centralized server environments.
• IBM Z runs 70 percent of business workloads at 10 percent of total IT cost – whereas distributed platforms handle 30 percent of workloads at 90 percent of IT cost.

The point here is that choosing to host workloads on mainframes sets you up for financial success by default. It’s not as if mainframes are hard to cost-optimize; on the contrary, mainframes are designed to save you money without compromising performance or reliability.

…But mainframe cost blunders abound

However, whether you actually yield a high ROI on your mainframes hinges on how effectively you leverage the unique functionality of mainframe systems.

Common mistakes that we see in the industry that undercut the value of mainframes include:

• Lack of cost visibility: This is probably the single biggest cause of low mainframe ROI. Businesses just don’t know how to calculate their mainframe TCO, so they lack a baseline for measuring and improving financial performance in the mainframe context.
• Underestimating software licensing costs: The greatest cost in most TCOs is software licensing costs, not mainframe infrastructure costs
• Unnecessary migrations: Sometimes, businesses decide they should migrate away from mainframes due to a perception that mainframes are outdated. This is a huge mistake in most cases not only because mainframes are not at all obsolete (on the contrary, as noted above, they continue to power a majority of global transactions) but also because organizations often underestimate the challenge of getting mainframe workloads to run well in commodity servers across all stages of the SDLC – dev, test and prod.

If you desire to get the most out of your mainframes, you need to keep your workloads running on them.  Rather than on more costly commodity servers – in ways that minimize your licensing costs and allow you to achieve continuous visibility into your mainframe spend and total TCO.

Greening the mainframe: How to optimize mainframe ROI

How, specifically, do you achieve those goals?

The answer starts with keeping your existing mainframe workloads running. Doing so avoids the major expense of migrating to commodity servers, as well as the cost of operating those servers – which, as we’ve noted, is much higher than mainframe operating costs.

At the same time, though, it’s critical to modernize your mainframe environments. Take advantage of technologies like z/OS Connect to run modern, API-driven applications on mainframes, or Open Data Analytics for z/OS to host modern data analytics workloads on top of your mainframe. You can also deploy Linux containers on your mainframe, allowing you to run distributed workloads on massively scalable and high-performing mainframe infrastructure instead of costly distributed infrastructure.

Through strategies like these, businesses can not only maintain, but actually increase the performance and scalability of mainframe systems while simultaneously reducing costs. They get the best of both worlds: Modern application architectures and deployment techniques combined with cost-effective, dense, ultra-reliable mainframe infrastructure.

Enjoy your green grass

Let me close with another analogy: If you find that your lawn isn’t as pretty as you’d like, you could go and pave it over with asphalt, or entomb it in astroturf. It wouldn’t exactly be an elegant solution, but it would save you from having to look at your yellowing grass.

That approach is akin to moving your mainframe workloads to commodity servers just because you’re not happy with the current state of your mainframe. It’s not really a good solution; it’s a knee-jerk reaction that may seem like a good idea in the short term, but that shoots you in the foot in the long term.

A better approach is to take the steps necessary to modernize your mainframe. Even if you’re not pleased with the state of your mainframe today, you can improve it – just like you can water and fertilize a stale lawn to bring it back to its prime