Enterprise Executive 2016: Issue 2 : Page 16
Overcoming Digital Denial: Accepting and Planning for the of the Mainframe FUTURE By Chris O’Malley reinvestment are not necessary. These mainframe ignorers and mainframe status quo’ers are both in deep denial. And their denial is profoundly toxic to the organizations they serve. In fact, the only way forward for IT as a whole is to soberly accept and plan for the future of the mainframe as we can reasonably expect it to unfold over the coming years. Readers of Enterprise Executive generally understand that predictions regarding the Why the Mainframe Will Persist | 2016: Issue 2 The 16 | tech sector as a whole remains in deep denial about the future of the mainframe. In fact, anyone who didn’t know better would think from reading most of what’s written about information technology that the mainframe has already ceased to exist. Just as, unfortunately, most of those who acknowledge and celebrate the ongoing role of the mainframe in IT give the impression that the platform’s technology, processes and people are pretty awesome as they are—and that an intense rethinking, innovation and E nt e rp r i s e E xe c u t i ve
Overcoming Digital Denial: Accepting and Planning for the Future of the Mainframe
The tech sector as a whole remains in deep denial about the future of the mainframe. In fact, anyone who didn’t know better would think from reading most of what’s written about information technology that the mainframe has already ceased to exist.
Just as, unfortunately, most of those who acknowledge and celebrate the ongoing role of the mainframe in IT give the impression that the platform’s technology, processes and people are pretty awesome as they are—and that an intense rethinking, innovation and reinvestment are not necessary.
These mainframe ignorers and mainframe status quo’ers are both in deep denial. And their denial is profoundly toxic to the organizations they serve.
In fact, the only way forward for IT as a whole is to soberly accept and plan for the future of the mainframe as we can reasonably expect it to unfold over the coming years.
Why the Mainframe Will Persist
Readers of Enterprise Executive generally understand that predictions regarding the demise of the mainframe are foolish and unfounded. But in case anyone needs a refresher, here are some key concepts to bear in mind:
• Investments in mainframe code are irreplaceable and not replatformable: Telling a digital enterprise to ditch its mainframe is like telling a person to get a brain transplant. Companies have spent years honing and refining their mainframe application logic to meet the needs of the business. There is no compelling reason to undergo the extreme expense and risk of attempting to replatform that application logic or associated databases.
• The performance economics of the mainframe are unmatched: While companies without mainframes chase cheaper performance at scale in the cloud, IBM z13 owners can execute 2.5 billion transactions—the equivalent of 100 Cyber Mondays—on a single machine every day of the year. And they can do it with far less OPEX than it takes to manage the typical hypercomplex x86 data center or, even worse, to pay the absurdly high and unpredictable monthly charges of today’s premium cloud providers.
• The mainframe is outevolving x86 hardware: The debut of the z13 last year demonstrated that—far from being a moribund platform—the mainframe is actually evolving more rapidly than distributed computing. Virtualization, software-defined everything and the move toward converged architectures in the x86 world are actually evidence that commodity infrastructure vendors now recognize the superiority of the mainframe model and are playing catch-up.
The mainframe’s existence, in other words, will remain a beneficial presence in IT for the foreseeable future. Any bet against its persistence will be a losing one.
Why the Mainframe Must Change
While the mainframe persists in some form or another, it absolutely cannot continue as it is today. Reasons for this include:
• The need for speed: The tools and processes most IT organizations use to manage mainframe development, test/QA and production ops are obsolete holdovers from a time when slow, waterfall methods were adequate for the business.
That is no longer the case. The pace of change in today’s highly volatile markets is unrelenting. So, if the business needs IT to update a mainframe application or extract a test data set from a mainframe database, it has to happen fast. Mainframe tools and processes—and perhaps even the underlying culture of mainframe teams— must be transformed to match the agility that has been achieved on other platforms.
• Cross-platform convergence: The days of the monolithic mainframe application are over. Increasingly, mainframe applications and data actually serve as back-end resources for customer-, employee- and partner-facing web and mobile apps that mix-and-match code of all kinds running on platforms of all kinds.
Mainframe teams can therefore no longer work in their own little silos. Instead, developers, test/QA staff and ops teams must be able to collaborate more seamlessly across platforms (mainframe, distributed, cloud, web, mobile, etc.) and disciplines (development, security, problem management, capacity planning, etc.) to optimally meet the escalating digital requirements of the business.
• Generational shift: Veteran mainframe developers, DBAs and operations techs can’t stay in the workforce forever. At some point, their work will have to be turned over to a new generation of IT professionals.
Some small percentage of those new IT pros will come out of specialized mainframe education programs that have been established at various colleges and universities. But most next-generation IT staff will be steeped in technologies like Java and Eclipse. They will have little if any mainframe-specific knowledge and near-zero tolerance for tools that are not intuitive, automated and graphical. Despite this radical shift in staffing, IT must come up with a way to keep updating its core COBOL applications, leveraging its DB2 data and maintaining performance of its z Systems environments.
In other words, IT must accelerate mainframe processes, integrate mainframerelated work into its broader enterprise DevOps workflows and make mainframe DevOps 100 percent millennial friendly.
Key Enablers of Mainframe Transformation
To get their mainframes from current state to required state, IT leaders need to make several specific changes:
• Adopt lean/agile processes: Most mainframe development teams are stuck in waterfall mode. This is in part a process problem and in part a tool issue. If your source code management system doesn’t make it easy to check work in and out—or, perhaps even more important, to assign work to developers in parallel—your turnaround times will never be as fast as the business needs them to be. An SCM upgrade alone won’t do the trick, however. You also have to get developers thinking in agile terms (scrum, Kanban, etc.) so, that work is driven by the urgency of the business rather than the habits and convenience of developers.
• Embrace smart visualization: The best way to make the mainframe millennial friendly is to adopt more visually intuitive mainframe DevOps solutions. Visualization entails more than just grafting a pretty interface onto the same old tired tools. Truly transformative visual tools allow developers, data analysts and other DevOps process participants to do their jobs quickly, without spending hours hunting through line after line of code or arcane program analysis. Instead, they can easily and immediately see how a program’s logic flows, how that program relates to data sources and other programs, where an external call in a program run-time is bottlenecking, etc. Visualization thus both enhances the productivity of mainframe experts and enables mainframe novices to successfully perform many tasks that are currently beyond their skill level.
• Drive de-siloed collaboration: As noted above, end-to-end applications increasingly include both mainframe and nonmainframe components. If those components are created by entirely siloed teams using entirely separate tools in entirely disconnected processes, the speed, quality and efficiency of application delivery will all suffer. IT leaders therefore need to promote collaboration among DevOps stakeholders across all platforms and technologies. This collaboration can be facilitated through a shared SDLC environment, department restructuring, team incentives and other means.
The Mainframe Transformation Leadership Checklist
None of this, of course, happens by itself. Mainframe transformation requires leadership. Before taking action on mainframe transformation, IT leaders should therefore perform a thorough and honest self-assessment. That self-assessment should include the following questions:
• Am I definitely free of mainframe denial? Again, that denial generally takes one of two forms. One is a deluded belief that the mainframe is somehow eventually just going to go away. The other is that the mainframe status quo is “good enough.” IT leaders who drift too far toward either of these states of denial cannot successfully engineer mainframe transformation.
• Do I have a clear vision of the mainframe’s future? It’s not enough to know that the state of the mainframe has to change. IT leaders need to know exactly how it has to change. That vision must encompass a future in which the mainframe:
- is highly agile
- exists as a “first among equals” platform integrated with distributed and cloud
- is staffed by DevOps professionals whose skills and careers are not mainframe centric.
• Am I willing to strategically reinvest in the mainframe? Transformation of the mainframe increases value and reduces costs. But getting there can require some incremental investment in new tools and training. IT leaders who focus exclusively on mainframe cost reduction—and refuse to engage in any smart reinvestment at all—will not be able to reap the substantial benefits and savings associated with agile cross-platform DevOps.
• Am I prepared to shift vendor partnerships? Mainframe transformation doesn’t just require new investments. In many cases, it may also require different investments. IT leaders must therefore be willing to rethink and realign even their longest-standing vendor relationships without the constraints often imposed by inertia or misplaced loyalties.
• Am I willing to lead up and down? At most companies, mainframe transformation doesn’t just require IT leaders to lead the people below them on the org chart. It also requires championing and advocacy up and across the org chart as well. That’s because the mainframe is not the “flavor of the week” like mobile or cloud—so, it may not be well accepted by the C-suite. IT leaders must be prepared for this executive resistance and have the force of argument to overcome it.
IT leaders who confidently check off all the boxes above can provide their companies with significant competitive advantage by transforming their mainframes into powerful, agile sources of digital superiority for years to come. Those leaders who can’t check off all the boxes need to seriously reconsider their thinking about the mainframe—because denial, delay and disinvestment are not viable IT strategies for an app-centric future.
Chris O’Malley is CEO of Compuware. With nearly 30 years of IT experience, he is deeply committed to leading Compuware’s transformation into the “mainframe software company for the next 50 years.” His past positions include CEO of VelociData, CEO of Nimsoft, EVP of CA’s Cloud Products & Solutions and EVP/GM of CA’s Mainframe business unit, where he led the successful transformation of that division. Email:firstname.lastname@example.org