For more than 50 years mainframes have powered thousands of organizations around the world, from banks to militaries to government agencies. Looking looking back at all that history makes me think about the critical role that big iron has played in the world, but it also gets me thinking about the future and what the next 10 years holds for the mainframe industry.
+ Also on Network World: The future of virtualization: Don’t forget the so-called 'old' +
What makes me so confident that mainframing even has a 2026 worth looking forward to? After all, hasn’t the cloud revolutionized data storage and processing and ushered in the end of mainframes? The truth is that not every disruptive development replaces what it disrupted—sometimes not immediately and sometimes not at all. Globalization did not kill American IT jobs, Metallica didn’t negate Van Halen, and the cloud won’t kill mainframes because mainframes have something that the cloud will need over the next decade: power.
“Power,” says politician Francis Underwood in the hit series House of Cards, “is the old stone building that stands for centuries.” And while nothing in computing stands for centuries, mainframes are the most powerful computers in the world—powerful enough to last at least another few decades. No other machines can match the raw processing muscle and efficiency of a mainframe, to churn as much data as these massive computers, and for this reason, I think the cloud will come to rely on big iron. As the world becomes more and more data-driven, the data will need more and more horsepower, and a mainframe is the perfect machine for the job.
The human race today produces more data every 14 minutes than is contained within the Library of Congress. That’s just today, with our smartphones and computers and basic wearables such as FitBits. Now raise your hand if you think we’re going to be producing less data by 2026. Nobody? Let me give you just two reasons why you’re right to keep your hands down.
What about the nascent Internet of Things (IoT), the uplink of everyday appliances to enhance daily life? Today some upscale cars, refrigerators and other major appliances have such functionality. But what happens when that technology gets democratized with a lower price tag? What about when your coffee maker knows what time you get up and your toaster knows how you like your toast? What about the amount of information it’ll take to run a nation’s worth of self-driving cars? To process that kind of data and do analytics in real time you need a lot of mainframes.
And I’m sure a lot of us remember mocking Google Glass just a few years ago. But Google Glass was ahead of its time. The Apple Watch is doing OK, but the bigger story is that the dawn of the smart wearable has just begun. In the next 10 years, our wearables will track our vital signs, moods, social networks, daily habits and careers. We’ll use them to play augmented reality games that make megahit Pokémon Go look like ET for the Atari. These watches and visors will be everywhere, tracking absolutely everything on the cloud. To process that kind of data, you need the kind of processing and analytics power that mainframes provide.
It’s not about cloud vs. mainframes. The cloud is not going to replace mainframes. In fact, we’re going to need mainframes just to churn all the data coming in from our smart wearables, our interconnected appliances and our robocars. Banks and other fiscal institutions will continue to want a massively scalable system that guarantees high levels of performance and data availability as well as world-class security. And as versatile programing tools let programmers code for mainframes without needing to know COBOL, big iron will continue to be a powerful engine driving our skyrocketing data usage.