Most corporate business processes and customer information still reside in large, complex COBOL applications and databases running on mainframe computers. These are often the forgotten systems of record in a datacenter -- forgotten because they are so reliable, efficient and secure that many people take them for granted.
These strengths are also why mainframe applications power the global economy and feed an increasing number of web and mobile applications (aka systems of engagement). For example, when a customer cashes a check with a mobile device, a mainframe is called upon to ultimately complete the transaction.
Now CIOs are realizing this symbiotic relationship can be leveraged for more progressive innovation. A Forrester survey we at Compuware Corporation commissioned found 96% of new business initiatives at large organizations involve mainframes.
But the progress of those initiatives is suffering because mainframe development teams often struggle to keep pace with web and mobile teams, ultimately reducing their ability to deliver the digital innovation customers crave.
What’s Constraining The Mainframe?
The struggle is unrelated to system hardware. The mainframe is one of the most modern IT systems available. The newest model, the z14, arrived in the latter half of 2017, and it drove IBM’s year-over-year mainframe hardware revenue up 71% last quarter.
The mainframe’s struggle is also unrelated to COBOL, the mainframe’s primary programming language. While some COBOL applications were written decades ago, these applications contain business processes that have been honed and perfected over time.
The real problem is that too many mainframe teams still adhere to a waterfall-based (linear and typically slow) operation. The front-end systems the mainframe increasingly feeds are surrounded by open and idea-driven cultures, iterative and collaborative processes, and automated, integrated tools. In contrast, waterfall upholds:
• Siloed development and operations teams, preventing collaboration.
• Slow and protracted development and delivery, preventing the timely release of innovation in a constantly shifting digital world.
• Outdated, esoteric tools that are often incapable of integrating with non-mainframe toolchains.
This regressive IT mode is incongruous with digital transformation. Enterprises must change their mainframe environments to leverage the platform’s unique strengths for continuous digital innovation.
Making The Transition From Waterfall To Agile
It’s time for enterprises to pop the mainframe’s build-build-build bubble that forms in waterfall and “shift left” toward more iterative development that is centered on customer feedback and dedicated to continuous improvement.
This will enable the continuous delivery of innovation, releasing in two-week-or-less cycles called sprints rather than focusing on one massive, complex deliverable that takes a year or more to release and often misses the mark on customers' ever-changing expectations.
Based on our experience in transitioning from a staid waterfall enterprise to an agile DevOps software company, there are three core changes enterprises should make to back the transformation of their mainframes into systems that keep pace with the digital initiatives they increasingly drive.
1. Commit To Comprehensive Change
Commitment to change must be inclusive and unconditional. When only one team changes, any chance to fault a new methodology will be taken, as waterfall may appear better only because it is more familiar. Approaching an all-in transition across development teams will remove the tendency to revert to waterfall and will lead people to learn the new way, helping teams move forward faster and recognize the benefits of change sooner.
Additionally, today’s enterprise applications are collections of applications that execute across several systems. Many enterprises assume they can innovate systems of engagement without touching the mainframe, and many mainframe teams believe they can gradually adjust routines they feel comfortable changing. Neither can go halfway with change. They should commit to driving the same velocity, quality and efficiency across all systems.
2. Equip Teams With Modern, Integrated Tooling
Too many mainframe teams use inadequate tooling based on an esoteric, keyboard-driven interface. Enterprises should equip mainframe teams with modern alternatives that provide automation as well as graphical views of programs and data to enhance understanding and productivity. These are also what computer science graduates are proficient with and expect to use -- whereas awkward, antiquated technology will drive them away.
These modern tools should integrate with each other but also with non-mainframe solutions to include the mainframe in a toolchain that spans across IT systems. This will close the gap between mainframe and distributed platforms for the seamless support of multiplatform projects between them.
3. Break Down Cultural Silos
As the previous steps are executed, the natural result is that mainframe teams will become more collaborative. But enterprises must continually work at pulling mainframe talent out of its silo, too, enabling mainframe developers to collaborate and ideate with non-mainframe developers as well as the broader IT operations team. This will make way for improved development speed and quality by introducing constant communication and feedback.
The mainframe has an invaluable and irreplaceable role in the digital landscape, but if its waterfall-based issues are left unaddressed, it also has the potential to become an enterprise’s biggest constraint to digital transformation. Using these best practices to transition your mainframe away from waterfall will help it evolve from a data-center pariah into just another modern system in the enterprise.
Chris O'Malley is president and CEO of Compuware, the world's leading mainframe-dedicated software company