Chris OMalley

By Chris O’Malley, President and CEO, Compuware

imgresConventional wisdom maintains that companies should abandon the mainframe and do all their computing in so-called “hybrid cloud” environments that combine virtualized x86 infrastructure in their own data center with the XaaS resources from cloud providers.

A sober examination of IT economics, however, leads to a very different conclusion. Yes, a hybrid IT environment is optimal. But the two platforms that comprise that hybrid environment should be the mainframe and the XaaS cloud.

It is, in fact, on-premise x86 infrastructure that companies should abandon as soon as possible—because it is both horrifically expensive and a significant impediment to business agility.

IT reality check

The excessive cost of commodity x86 infrastructure is well documented. IDC, for example, shows server opex spiraling out of control—even as capital spending on servers remains flat. Rubin Worldwide has also documented the high costs borne by server-intensive IT organizations, as compared to more mainframe-intensive ones.

Every IT professional’s own experience bears this out. IT spends countless hours patching, upgrading, and troubleshooting the almost insanely complex conglomerations of servers, storage devices, network devices, operating systems, and other data center detritus they have accumulated over the years. It takes squadrons of experienced, well-trained technicians to keep these hyper-complex environments afloat. And they simply cannot be secured—because the threat surface and threat matrix are always expanding faster than enterprise defenses.

Perhaps worst of all, the incremental costs for new applications and services are huge. More code requires more infrastructure, which requires more staff and more data center capacity to house, power, and cool. So there are minimal economies of scale.

Contrast this with the mainframe—which delivers exceptional performance, unmatched security, rock-solid reliability, and incremental data center costs that approach zero—and the economic logic of two-platform IT quickly becomes evident.

The mainframe thus trumps commodity on-premises infrastructure when it comes to running applications—such as product development and customer analytics—that provide your company competitive advantage

Everything else—HR, office productivity application, sales automation, and the like—can run in the cloud. PowerPoint and payroll don’t deliver competitive advantage, so it makes sense to run them where economies of scale are greatest and capex is non-existent. Plus, if a better solution comes along, you can readily change your XaaS provider.

So why not?

The logic of two-platform IT is irrefutable—which means there are only three reasons IT organizations aren’t doing it more:

  • Many IT leaders still choose to pretend that the total cost of ownership for distributed infrastructure is reasonable. But it’s not. A full accounting of all IT-related operating costs will bear this out.
  • A lot of people—including IT staffs, IT vendors, and IT contractors—have a personal stake in the perpetuation of the data center status quo. Two-platform IT requires leadership that puts the interests of the company ahead of the interests of IT practitioners.
  • The dominant culture as promulgated by analysts and the tech media has been staunchly anti-mainframe and pro-commodification for years. It takes a certain amount of courage to declare that the Emperor has no clothes—and to act accordingly.

Despite these hurdles, two-platform IT is an immensely practical strategy that can yield tremendous gains in both economy and agility. It’s not for the faint of heart, and it can’t be done without committed leadership. But you can run your company without a data center that looks like a hairball from cyberhell.

I know, because we’ve done it at my company. And we’re very glad that we did.