The consensus is clear and, now, strongly supported by both Forrester and Gartner. Enterprises must bring agility to mainframe development. After all, large enterprises run on mainframe code and mainframe data. So it’s imperative that they be able to nimbly twist that code and data into whatever shape is needed at any moment to rapidly achieve digital market advantage.
In fact, anything less than full mainframe agility will doom old-line businesses to market irrelevance.
It’s not enough, however, for enterprise leaders to be convinced that it’s time to make their mainframes agile. They need to know how.
Here, then, is an overview of the best-practice process by which mainframe environments are being transformed from their traditional a-couple-of-code-drops-per-year-if-we’re-lucky state to full card-carrying membership in the Agile enterprise.
First, modernize the work of mainframe coding to increase developer speed and productivity without compromising quality. Writing mainframe code isn’t actually that difficult. What’s difficult is using painfully-outdated, mainframe-esoteric tools to work on mainframe application logic and data dependencies that are highly evolved and poorly documented. So mainframe agility first requires new tools that make it easier for veteran mainframe programmers—as well as Millennials with minimal mainframe experience—to do more work in less time while still delivering code of esoteric quality.
Next, include the mainframe in your broader adoption of Agile/DevOps management disciplines. Once you’ve addressed mainframe developer productivity and QA, the next constraints to attack are those that prevent those developers from tackling coding tasks with appropriate customer-driven urgency. To achieve this objective, you need scrum management that empowers teams to work in parallel across platforms without stepping on each other’s toes—and that enables you to more nimbly re-prioritize their collaboration based on ever-changing business and operational inputs. The is an essential phase in transforming from a mainframe siloed culture stifled in maintenance/projects into an IT inclusive culture inspired in innovation/experimentation.
Finally, make mainframe release management more automated and more intelligent. As you streamline “dev,” your next constraints will be “ops”—which means you’ll want to streamline release management across your mainframe and non-mainframe environments. This coordination will ensure that you can rapidly and frequently release new mainframe code in support of your rapid and frequent release of new customer-facing apps. Your ability to confidently accelerate your release cycles also depends on your ability to automatically and safely roll back those releases if and when necessary.
In most cases, enterprise IT will break these three phases of mainframe transformation into several even smaller steps. That’s wise, because “boil the ocean” initiatives rarely bear much fruit. Plus, a staged approach delivers tangible near-term wins that pay off economically while helping you build buy-in for your long-term goal.
Shameless plug: Compuware and our partners have solutions for successfully executing on the process described above. These solutions include:
- Topaz Workbench, Program Analysis and Enterprise Data within Eclipse for graphical analysis of mainframe programs and data dependencies—along with SonarQube, SonarLint and Jenkins for on-the-fly insight into code quality issues.
- Our recently acquired ISPW technology with Jira for issue tracking and Confluence for team collaboration.
- ISPW Deploy combined with XebiaLabs’ XL Release for cross-platform release automation.
Of course, this entire process presupposes strong executive leadership. Without that leadership, transformation efforts will likely be undermined by recalcitrant mainframers bitterly resistant to change and woefully inadequate financial investment in, essentially modern, mainframe development tools.
With the right leadership and process, on the other hand, your mainframe can be radically transformed to a fully Agile platform in a fully Agile enterprise. And that’s a must for any large enterprise hoping to remain relevant in marketplace of relentless digital innovation.
Originally Published on LinkedIn Pulse.
- CIO Challenge: Dodging the Bimodal IT Bullet - Mar 13, 2019
- Yes, There Is a Clear Path to Mainframe Agility - Dec 26, 2018
- Digital Denial: Why You Can’t Trust the Application Economy - Jul 4, 2018