For years, mainframe professionals have endured pervasive unfounded rhetoric about how theirs was a “legacy” platform, doomed to irrelevance—and, ultimately, demise. At Compuware, however, just the opposite is true. It is the insanely complex and costly x86 commodity server infrastructure hairball that we have consigned to the scrap heap in favor of a two-platform IT strategy based exclusively on mainframe and the cloud.
Ending the madness
Our reasons for banishing x86 servers from our datacenter are obvious to anyone who has ever worked in corporate IT. A computing environment built from lots of separate pieces of hardware acquired from multiple vendors over multiple years inevitably becomes an opex nightmare. Some component somewhere is always failing or getting overloaded. OS patching and re-configuration is an endless task. And the threat vulnerabilities are so vast that it is for all intents and purposes impossible to defend or keep in full compliance.
Worse yet, economies of scale are negligible. Every time you add applications, services, capabilities, users, or capacity, you add cost. And that cost invariably includes more staff—because complexity eats labor for breakfast, lunch, and dinner.
IDC, Rubin Worldwide, and others have published plenty of data demonstrating just how horrifically inefficient commodity distributed infrastructure is to operate. But two recent developments in the x86 market are particularly damning. One is that a single company now rakes in $6 billion annually just to do to x86 compute capacity what we’ve been able to with mainframe compute capacity for decades: virtualize it. Even more outrageous, that single company’s revenue is greater than that of all independent mainframe software companies combined. That’s a stunning indictment of x86 opex.
The other damning piece of evidence is that the leading x86 server vendors are now hawking something they call “convergence”—which is essentially a desperate attempt to market x86 servers in a package that looks and acts like a tiny, poorly architected mainframe.
In other words, the x86 world has basically surrendered and agreed that the mainframe way of doing things is better. The only question is why anyone would go with an inferior imitation when you can have the real thing?
The Cloud as Complement
This is not to say that there is no place in our organization for x86 machines. There is. That place is the cloud. If a SaaS provider has invested in a purpose-built, highly homogeneous x86 environment in order to deliver a SaaS solution—and if that SaaS provider’s business model makes it profitable for them to manage that x86 environment for us—so be it.
We are happy to rent workloads that don’t provide us with any competitive advantage from such a partner. There is no reason for us to invent, deploy, and manage our own systems for payroll or sales automation. So we consume them as services from the cloud.
On the other hand, there is digital intellectual property that differentiates us. This includes our product source code and confidential information about our company and our customers. It also includes proprietary Linux applications that are integral to our customer engagement and to our innovation agenda.
This digital intellectual property belongs on the mainframe. The mainframe is highly reliable, highly scalable, and highly secure. It delivers unmatched performance. And when we add new applications or new analytic capabilities to our mainframe environment, we add little or no incremental costs. We don’t have to hire more people. We don’t have to consume more power or add rack space. So we can grow and change our business efficiently, economically, and without untenable complexity.
Essentially, we are creating our own internal mainframe cloud—making greater use of the mainframe’s power and efficiency through virtualization.
Is Two-Platform IT for You?
The advantages of the mainframe and the shortcomings of x86 chaos are probably not news to regular visitors to this site. What probably is news is that a company like ours has fully opted out of on-prem x86 infrastructure and gone all-in with the mainframe as our platform-of-choice for strategic workloads—while embracing the cloud for on-demand access to commodity services.
But we believe our two-platform strategy makes sense. It reduces costs, optimizes our agility, and keeps our IT operations reliable and secure.
If you’d like to hear more about how and why we’re doing two-platform IT, check out our YouTube video. And let us know if you’re thinking about moving to two-platform IT too. We’d be glad to share our lessons learned thus far with you.
Editors Note: Thanks to Chris O’Malley the CEO of Compuware for these insightful words. Chris is a great supporter of the mainframe and his passion for the platform is helping to drive the #mainframedebate…