The world still runs on mainframes

One might be tempted to assume that the march of technology always means “out with the old, in with the new.” But what if the so-called “old” technology is responsible for processing 87% of all transactions in the world; is favored by 44 of the world’s top 50 banks and the vast majority of top retailers; and already affects our lives every day? We’re talking about mainframes, such as IBM Z, which have hidden in plain sight and powered the world of commerce since the 1960s. (In fact, the IBM System/360 mainframe figured prominently in a Mad Men episode.)

Mainframes are as misunderstood as they are ubiquitous. “If you used a credit card today, you used a mainframe,” said Tina Tarquinio, Chief Product Officer for IBM Z and LinuxONE, as part of a panel called “Mainframe Mythbusting” for New York Tech Week. “If you booked a flight, refilled a prescription or pulled cash from an ATM, you used a mainframe.”

So what do these silent heroes do? As IBM’s PJ Catalano, Principal Test Lead for IBM Z and LinuxONE (and mainframe influencer), told IBM Think in an interview, “Mainframes are big computers that are great at processing a lot—and I mean a lot—of transactions at scale, with both speed and reliability.”

He gave an example of how the user-end experience of a seemingly simple bank transaction is just the tip of the iceberg: “When I go check my online bank account, I only think about my single set of actions: log in, check a balance on one account, maybe transfer some money to my kids’ accounts and log off.” That single action involves 5-10 interactions with the bank mainframe, Catalano said. Multiply that by, say, 10,000 customers transacting every second, with fraud verification added on top of that, he explained, and you begin to understand the sheer volume of information and processes that customer activity requires.

“That is what mainframes are built for: doing the work that powers the economy, powers the travel industry, powers healthcare systems, each and every day,” he said. “And that’s why they deserve lots of hugs.”

Still, a lack of public awareness about mainframes might be contributing to a number of myths about the technology, the “Mainframe Mythbusting” panelists said, including the notion that mainframes are being phased out because younger generations don’t use them. Not true, said Tarquinio, who pointed out that even modern tools like Python, Go, Visual Studio Code and Git all run on the IBM Z ecosystem. And anyone using those tools is also interacting with a mainframe, whether they know it or not. “All of the ways that people are learning to program today exists on the [mainframe] platform,” she said from the floor of the IBM auditorium, “and [the new formats] are designed to interact seamlessly with older code.”

Another common misperception, IBM Staff Writer Aili McConnon and moderator of the panel shared, is that “if you want to be cutting edge, you have to rip up and replace the mainframe with something else.” But this is a move some enterprises will likely come to regret, she said, citing a Gartner study reporting that “by 2030, 75% of vendors operating in the ‘mainframe exit’ market will either pivot their business models or cease to exist.” The reason for the exit failure, according to the report, is that “for most large-scale enterprises, the sheer volume and interconnected complexity of this data make wholesale migration a physical and financial impossibility.”

To illustrate just how unwieldy a mainframe exit can be, Tarquinio offered an analogy. “Imagine that someone said the way to modernize the world would be to replace your phone with another phone. And that you have to move everything over to the new phone, app by app. Painful,” she said. “Now, imagine that [one of those apps] was the core app that ran your business, and how risky that would be.” Migrating from a mainframe platform to a piecemeal substitute would be at least that chaotic, Tarquinio said. Meanwhile, mainframes are designed “so that you don’t have to make that big leap,” she explained. “Wherever you are, you can modernize and have everything interact the way they need to, without risking everything.”

Some mainframe experts fear that enterprises might be underestimating the role that a large volume of information plays in the complexity of a system—something that mainframes were born to handle. Echoing the Gartner report, Catalano noted, “Mainframes give you a simple, unified approach to managing work at scale. Take the same amount of work and distribute it across thousands of small servers, [and] that just opens up a whole host of problems.” Moreover, curveballs like security updates and network configurations can “disrupt thousands of systems,” he said. “On the mainframe, it’s one-stop shopping—one place to touch, one update, one configuration change. And all of it can be done concurrently without impacting the work that is happening.”