Chris O’Malley is CEO of Compuware.
Over the last year and a half, one of the few bright spots in IBM’s search for sustained revenue growth has been the z14 mainframe, which has achieved the strongest mainframe performance in decades following IBM’s highly successful z13 cycle. But, are the last four years of strong mainframe performance an aberration or the beginning of an ongoing trend?
BMC’s 2018 mainframe survey results show 92 percent of respondents predict long-term viability for the platform—the third straight year that number has increased. Furthermore, IDC recently found 81 percent of its clients plan a repatriation of workloads from public clouds to on-premises private clouds, hosted private clouds or other on-premises infrastructure (including mainframes), with this figure expected to rise to 85 percent in 2019.
The mainframe has endured because it’s the ideal back-end system for managing proliferating workloads—it can process over one million transactions per second, largely fueled by the mobile megatrend. This is due to the platform’s superior reliability, availability, security and cost-effectiveness.
As reliance on the mainframe grows, the business demands on it are changing. Development and operations teams are now tasked with transforming the mainframe into an engine for agility, where it has traditionally been a siloed platform with slow processes by design that prioritized system reliability over customer value.
Enabling this massive change from a status quo mindset to one of continuously improving through agility requires transitioning to what we call the “mainstreamed mainframe.” Some of the keys to this transition are as follows:
1. Commit to a well-defined mission: The first step is to inspire with a customer-centric mission and then make sure everyone is on board. Many stakeholders—from operations staff to executives responsible for budget decisions—may have succumbed to popular misconceptions that an all-in move to the cloud is the only way forward.
To ensure buy-in, organizations must first make it clear to key stakeholders that the mainframe is here to stay—and with good reason: recent history has shown (4) costly re-platforming of the mainframe’s core systems is both impractical as well as a poor use of an organization’s finite financial resources.
2. Establish and consistently measure against KPIs: Traditionally, key performance indicators (KPIs) for mainframe operations have focused on metrics like “five nines” availability and time windows for batch completion. These metrics are useful and important, but sole reliance on them is outdated.
The next step is to institute a focus on continuous improvement in software delivery. This means mainframe developers’ and operators’ efforts should be consistently evaluated and held accountable to the same improvement mandates governing other platforms.
In BMC’s survey, 59 percent of respondents reported an increase in mainframe application transactions in 2018. No matter how sophisticated or advanced an application’s front end may be, poorly performing code on the back-end mainframe will result in a poorly performing application overall.
With the mainframe playing such a critical role in modern digital deliverables, it becomes imperative to consistently measure and improve mainframe software delivery velocity, quality and efficiency. A trade-off mindset (e.g., increased velocity will result in decreased quality) is a digitally fatal mindset. Continuous improvements in all these areas are the key to digital success, and no single platform involved in the overall effort can be omitted. Mainframe users should be able to gather and analyze behavioral data at their disposal to identify areas for improvement.
3. Foster a culture of collaboration: The mainframe’s isolated culture has become increasingly entrenched over the course of multiple decades. Mainframe IT’s ongoing allegiance to siloed work is a significant risk as organizations strive to deliver applications spanning multiple platforms with greater quality, velocity and efficiency.
For ops teams specifically, tools for the new generation of system stewards must be easier to use. They must clearly and intuitively highlight emerging trends in the mainframe environment and help operators make smarter decisions about resource allocation and batch job queues for intense and critical workloads.
Digital excellence demands open, non-territorial collaboration between developers and mainframe ops teams. Agile development on the mainframe, in particular, creates many useful opportunities for ops staff to participate in scrums, sprint demos and the like. However, fostering better collaboration also requires an accompanying cultural shift—creating an environment that banishes blame, encourages transparency, celebrates learning and is focused on eliminating constraints.
And perhaps most importantly, digital excellence also demands ongoing customer collaboration. Customers, always beautifully and wonderfully dissatisfied, should be your “true north” in setting direction.
4. Embrace automation: Increased automation, in both development and operations, is essential for mainframe agility. Thoroughly automated processes, particularly in the area of testing, are critical to enabling faster, more frequent promotion and deployment of high-quality code.
It’s surprising that mainframe code—the mission-critical heartbeat of many modern applications—is often omitted from automated processes applied to the rest of the DevOps toolchain. Only by implementing automated processes uniformly across all platforms involved in a DevOps effort can organizations realize the full benefits. Otherwise, there will always be a constraint in the chain, impeding the end-to-end code promotion effort.
For mainframe ops teams specifically, automation can insulate staff from underlying complexities, so they can continue to scale their mainframe environments as needed while minimizing unnecessary costs and keeping skilled headcount requirements down. One example is automated scheduling of batch jobs in a cost-aware manner. This helps save money on mainframe licensing costs (MLCs), which have the potential to rise significantly as transaction volumes increase.
5. Sharpen “find and fix” capabilities: Today’s operations teams assume the onerous responsibility of trying to tame an IT monster with multiple growing tentacles. As infrastructures grow more complex to support more users and computing workloads, the surface area for problems that can impact application performance naturally expands.
Precise mainframe diagnostic and optimization capabilities are becoming especially important as mainframe applications are increasingly called upon to provide back-end processing for customer-facing web and mobile services; and as those customer-facing services are updated more frequently in ways that directly impact calls to mainframe databases and applications.
With more effective tools and a new, expanded mission to accelerate business and collaborate more with development, mainframe ops teams must be empowered with accurate, more comprehensive information. Only then can they aggressively monitor their environments to find and fix issues impacting overall digital service performance and cost.
For example, equipped with the right insight, teams can search out and identify SQL statements and Db2 system services causing excessive CPU utilization, application wait times or application outages. They can even pinpoint specific resources, procedures and statements that are causing these issues—right down to the source code.
6. Utilize the cloud where appropriate: Finally, while the cloud may not be as qualified as the mainframe to serve as your back-end transactional workhorse, it’s an essential piece of the puzzle. Many organizations are finding tremendous success with a two-platform IT approach consisting of cloud services and on-premises mainframes working in tandem, with each platform supporting those applications and components for which they are best suited.
Conclusion
The mainframe isn’t going away for all the right business reasons, but it is being forced to change. Organizations and their IT teams must make the mainframe an equal citizen to other platforms within their infrastructures. Mainstreaming the mainframe through the modernization of culture, processes and tools is key. The steps described above are essential to advancing mainframe agility so organizations can continue leveraging the platform’s irreplaceable attributes.
Opinions expressed in the article above do not necessarily reflect the opinions of Data Center Knowledge and Informa.
Industry Perspectives is a content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating