A Window on the Future of Enterprise IT: Simplification, Innovation, DevOps
Overview: Learn how Compuware’s enterprise IT team has transitioned from maintaining servers in a data center to helping our company and its customers transform through simplification and innovation under Two-platform IT.
The Compuware of today is a far cry from what we were four or five years ago. We’ve transformed from waterfall to Agile, and we’re on a DevOps journey of continuously improving velocity, quality and efficiency across the company every day for customers.
A major component driving our successful and ongoing transformation has been a data center infrastructure strategy we call Two-platform IT. Through this, we’re prioritizing workload allocation to 1) mainframe computers and 2) XaaS resources from cloud providers, thereby eliminating the need for expensive, complex, on-premises x86 commodity server infrastructure.
In an effort to continually remove old equipment from the data center and simplify everything, in just four years since starting Two-platform IT we’ve eliminated 60 to 70 racks, with only a few left that are queued up for removal.
On an episode of The Modern Mainframe, Compuware’s podcast, Tasktop CEO Dr. Mik Kersten did a great job portraying this enormous transformation we’ve undergone after seeing it firsthand during a visit to Compuware’s Detroit headquarters with DevOps luminary Gene Kim.
“We saw the Systems z running just massive and high-transaction loads of COBOL and then behind them … we saw this graveyard with these tombstones and what looked like these police outlines of the dead bodies that were actually racks … I thought the most brilliant thing was that in each of those outlines of the dead racks, we saw the name of the system that was retired.”
If only we had a before-and-after photo… Imagine your typical data center, seemingly endless rows of server racks and webs of cables everywhere. Here’s what ours looks like today.
“What Chris and the Compuware team did was absolutely eye-opening, both in terms of the transformation that’s been happening over the past few years at Compuware … but also what it means to all those large organizations, many of whom are our customers at Tasktop as well, in terms of how their data center, their software stacks, how their mainframe investments are going to change over the course of the next year or two.”
The Impact of Change
Interestingly, in a following episode to Mik’s, notable DevOps thought leader Gene Kim expressed his curiosity as to whether the team—my team—managing this former hardware empire felt any sense of longing for how things once were.
The feeling of ousting all these servers is, admittedly, a mixture of joy and sadness. I have memories of months of effort spent on projects building and maintaining servers and all their complexity. Now we’re recycling them.
But this feeling is the natural outcome of change, like when you’re cleaning out an attic or a basement and haven’t looked at something in 20 years. It’s nostalgic, but it feels good to get rid of it, too.
Just the physical footprint of one rack—data center costs, overhead space, power, cooling—costs about $12,000 a year. The other cost is maintaining what’s on those servers, not to mention the expense of disaster recovery, which is critical when maintaining physical servers to accommodate for recovering a system elsewhere if the original is destroyed in, say, a building fire.
Our enterprise IT team understands the effort and cost that went into building our former server empire—most of us have been here 15 to 30 years—so we appreciate the savings we’re creating as we pull this hardware out and redefine our roles in the company.
A New Day in the Life of Enterprise IT
We still do things to support what’s left in the data center, but now we’re extending it to the cloud. If we can’t, we simplify where we can, such as by recently consolidating three NAS boxes with a footprint of three full racks into one smaller rack-mounted box the size of a small server. But the days of maintaining and supporting racks upon racks of x86 infrastructure are over. Not only has Compuware transformed as a result of this, so have our jobs.
Eliminating complexity is critical to adequately serving our “always beautifully, wonderfully dissatisfied” customers, as Compuware CEO Chris O’Malley often quotes Amazon CEO Jeff Bezos. Complexity creates bottlenecks that diminish our velocity, quality and efficiency in developing and delivering a solution.
For instance, in the past, developing a unique feature or function to meet a niche customer need sometimes required we spec out and purchase special servers, which we first had to secure funding for, then install. This put delivery of a solution six months or more out from the time a customer request came in. Often, we would find out a customer no longer needed our solution because they had found another in that time.
Eventually, developers would get impatient—knowing the delivery of the solution would take six months—and circumvent this bottleneck by building their own environments, resulting in shadow IT. This brought its own complications, of course—lack of transparency, security and privacy risk, auditing difficulties, more bottlenecks.
Today, leveraging primarily the mainframe and the cloud with Two-platform IT gives developers, customers and IT less complexity to worry about. We gain speed through virtually an infinite amount of resources we can use to spin up projects without worrying about disk space or complex licensing or how it affects disaster recovery.
The DevOps journey we began at the same time we embarked on our Two-platform IT journey is also helping. Whereas transparency formerly was lacking, in a DevOps culture, developers come to our area frequently to collaborate with us face-to-face on finding the best ways to accomplish things. When it’s time, we provision what they need with more simplicity and speed than ever before.
Simplification is key to enterprise IT manageability and empowering IT to contribute through innovative ways. Today, our team is spending most of our time adding new and improved value to Compuware’s business and customers.
For instance, network engineers like Matt Luchi are learning new skills to enable innovative solutions like Compuware Topaz on AWS, zAdviser and Test Drive. By making the infrastructure available, we’re empowering Compuware developers to focus on innovating for customers rather than worrying about maintaining environments.
We’re also constantly tackling Jira tickets to reduce bottlenecks for developers who are innovating daily on an iterative path to releasing major product enhancements or new functionality for customers every 90 days.
Refreshing servers and disk arrays, can take months. It’s very labor-intensive. But this new kind of work is labor-saving and gives our team—and our company—more capacity to invest intelligence and energy into solving problems that help advance our business and improve outcomes for customers.
As Mik said, this is “what the future of … organizations’ data centers and IT stacks looks like. And it’s very different than what it is today. But Compuware has actually taken that step already and is actually sitting there already to be viewed by anyone who’s interested.”
Watch this video if you’re ready to see the future of enterprise IT.