Jan 29 ’14
The Enterprise IT Performance Maturity Model: Embracing the New Normal of Mainframe
Alan Radding, a 20-year IT industry analyst and journalist, recently collaborated with Compuware Corp. to create a new Enterprise IT Performance Maturity Model. The model helps IT organizations gain maturity in the current IT landscape. Enterprise Executive sat down with Radding and Dennis O’Flynn, director of Product Management for the Compuware Mainframe Business Unit, to discuss industry trends and the new maturity model.
Enterprise Executive: Paint us a broad picture of the current enterprise IT landscape.
Radding: Today’s enterprise IT clearly isn’t your father’s mainframe. Your data center team probably never thought they would be processing mainframe transactions from mobile phones—or running analytic applications against unstructured social media data. And your development team likely never imagined they would be architecting compound workloads across the mainframe and multiple distributed systems running both Windows and Linux. Even your CIO is surprised to find your mainframe is responsible for serving up millions (or even billions) of customer-facing transactions a day.
O’Flynn: Perhaps the most disconcerting change for traditional mainframe shops is that need to accommodate distributed, open systems (systems of engagement) alongside the traditional mainframe environment (systems of record). And with that comes the need to bridge the gap that has long existed between mainframe and distributed teams. In the “new normal” of mainframe, navigating that divide is no longer an option—it has become a business imperative. But how do you do it?
Enterprise Executive: When you say the new normal of mainframe, tell us what you’re talking about.
Radding: The new normal of mainframe basically encompasses a set of new attributes. Mainframe applications that once served only back-end systems are now customer-facing. In addition, complexity continues to rise as more compound and composite workloads span multiple platforms and operating environments. As complexity rises, so do mainframe transactions and MIPS.
Simultaneously, those with the most mainframe experience and knowledge are typically baby boomers who have already begun—and will continue to—retire over the next 10 years.
Enterprise Executive: Let’s talk about the business impact of this new normal of mainframe.
Radding: Mainframe economics have definitely changed over the last few decades. On one hand, things have improved due to the steadily dropping cost per MIPS, as well as the introduction of power-packed, low-cost new mainframes such as the zBC12. And though mainframe data centers have always been under pressure to rein in costs and use MIPS efficiently, these pressures are increasing exponentially as businesses rely not only on the mainframe—but on IT as a whole—to contribute to achieving business objectives.
On the other hand, there’s increasing pressure to streamline operations due to the rise in quantity and type of mainframe workload. Transaction volumes are soaring as new mobile users generate more transactions, 24x7. At the same time, overnight windows for transaction processing are shrinking or disappearing altogether.
O’Flynn: Similarly, the demand for high quality of service (QoS) levels continues unabated. While the mainframe’s QoS remains at historically high levels, the demand for QoS now extends to compound workloads that cross mainframe and distributed environments and leave IT needing to meet new end-user experience performance expectations.
The mainframe should no longer be viewed as just a cost center to the business that needs to be closely monitored, but rather as an integral part of the application delivery chain as well as a revenue generator—and must be managed as such.
Enterprise Executive: So what’s a maturity model?
Radding: Maturity models help organizations improve processes amid change. The new Enterprise IT Performance Maturity Model can help IT organizations improve processes for managing application performance and mainframe costs as distributed and mainframe systems converge.
What you want in such a maturity model is something that recognizes the skills and culture gap and helps close it. It also incorporates new mainframe roles and workloads alongside open systems, cloud and mobile, while encompassing new tooling to address management and operations in this new environment.
Enterprise Executive: How is maturity broken down in this model?
Radding: The new model defines five levels of maturity and incorporates distributed systems alongside the mainframe and recognizes the new workloads, processes and challenges that will be encountered. At each level, it identifies five maturity categories. There’s application technology, mainframe attributes, culture, performance technology and, finally, process.
Meanwhile, there are five maturity levels, ranging from ad hoc—in which the mainframe runs core systems and applications, or the traditional mainframe workloads, with a green-screen approach to mainframe computing—all the way to business revenue-centric, meaning that business needs and end-user experience are addressed through interoperability with cloud and mobile systems and real-time analytics to support revenue initiatives.
Enterprise Executive: And where do most organizations fall in the model?
O’Flynn: Most IT organizations will likely find themselves straddling different maturity levels for the various categories. For example, although many have achieved levels 4 and 5 of application technology, culture and processes remain at levels 1 or 2. These disconnects mean IT still faces many obstacles and isn’t reaching optimal levels of service delivery and cost management. And this doesn’t just impact IT. There can be business ramifications, such as decreased customer satisfaction and slower revenue growth.
Enterprise Executive: What’s contributing to that gap you described?
Radding: While there are many factors, the largest appears to be that most organizations approach the mainframe and distributed environments as separate worlds. Given the interrelated nature of today’s enterprise, this approach won’t work going forward. Mainframe and distributed teams need a shared view of IT and must be able to communicate on the same level.
Enterprise Executive: In addition to the issue of mainframe and distributed integration, enterprise applications continue to grow in complexity, likely to span multiple tiers, platforms and environments, while involving multiple technology silos in management and support. With that in mind, how do today’s IT organizations troubleshoot enterprise application performance problems? And how do they manage mainframe resources to ensure efficiency?
Radding: In many cases, not very well. Days, even weeks, can be spent in war rooms trying to answer questions such as what exactly is happening? Where did the transaction originate? Whose fault is it? Typical answers include: ‘All my systems lights are green. The CICS transaction is functioning as designed.’ And possibly the answer most indicative of the problem, ‘I don’t care what’s happening over there; my piece of the application is working fine.’
There’s little room for effective collaboration when teams are viewing problems independently. In order to address that deficiency, organizations need a unified, single view of the enterprise—the entire enterprise. Even though today’s applications span multiple system environments, most application performance management (APM) tools focus on a single platform. So, different teams run different tools to address the same performance problems. Plus, numerous workarounds are needed to ensure information is collected, analyzed and shared among teams. An effective APM strategy would eliminate this inefficiency and remove the gap between mainframe and other non-mainframe teams.
O’Flynn: It’s also important to see how a transaction flows and interacts. But the effectiveness of tracing compound business transactions end-to-end fades as the transaction components extend further, crossing multiple systems and management domains. A single, unified view would aid this process.
Finally, it’s necessary to understand the business context of a transaction. Similar to not having a unified view of transaction performance, not knowing the business context of a transaction greatly hinders collaboration. Once a CICS transaction hits the mainframe, the mainframe team can only analyze if each individual transaction is optimized. They need to see what was generating that transaction.
Radding: All these problems hold IT organizations back from achieving higher levels of maturity and realizing the full potential of their mainframe. Bridging the gap and smoothing cultural differences between various technology disciplines will provide a starting point for processes and service delivery improvement.
Enterprise Executive: The first challenge IT faces in almost any undertaking is a resistance to change. Apart from that, what other challenges need to be overcome to achieve mainframe maturity?
Radding: There are a few. One significant challenge is changing skills. And that’s not only about the retirement of skilled mainframe veterans, but now there’s also a need for service-oriented architecture (SOA), cloud and mobile technology skills. The demand for traditional mainframe skills isn’t going to diminish even as organizations seek to cultivate those new skills. To address these issues, it’s important to start hiring and training replacements now. Consider cross-training developers so you can impart mainframe experience to your distributed teams and vice versa.
Enterprise Executive: Any others?
O’Flynn: Yes. There are also things such as management visibility, end-user engagement and culture conflicts.
It’s important that management has measurement capabilities and increased visibility into the expanded IT infrastructure. And equally important are the users. End-user experience needs to be a central element of every technology decision your organization makes. And finally, bringing teams together. By having shared tools that offer a complete view across mainframe, distributed and mobile/cloud environments, you can reduce conflict and start troubleshooting issues more quickly and efficiently.
Enterprise Executive: So, it seems organizations that move up the enterprise IT performance maturity model will be well-positioned to compete in the continually evolving business landscape. What are some key takeaways for an organization seeking to achieve that maturity?
Radding: I’d say there are four areas of focus to keep you on the right path. First, address the skills gap and workforce churn by employing modern mainframe tooling. Second, expand your APM capabilities for multiplatform, end-to-end visibility. And third, navigate cultural divides by continuously engaging multiple organizational units. Finally, you need to leverage your technology to work toward continuous mean time to repair (MTTR) process improvement.
The Enterprise IT Performance Maturity Model can help organizations keep pace with the demands of customers and the business. Mainframes really can be more than just legacy systems; they can be engines for driving optimal business growth and success.