Forrester Consulting’s Research study “The Mainframe Opportunity—IT Strategies for Achieving Breakthrough Value,” which was published on behalf of CA, makes interesting reading. The report was based on detailed interviews with CIOs at large mainframe user organizations in North America and Europe.
In many ways, the results are as expected. The executive summary concludes that the mainframe is the most efficient, reliable, and scalable platform available; the System z’s operation is complex and requires specialized knowledge, provided by an aging and diminishing population of suitably qualified professionals; the perceived skills shortage is holding back further mainframe development and even encouraging users to move work off the mainframe and on to distributed systems; and the answer to this problem is a consolidated system management solution with an intuitive GUI.
I can’t disagree with any of the above. Of course, it’s also true that many mainframe installations have been trying to implement a single, consolidated, intuitive management system for decades now and it has proved to be a moving target. I’m confident, though, that many are now much closer to achieving this objective than they were a few years ago.
The user quotes really highlight some of the reasons why the System z has survived so many IT trends and revolutions. For example, a CIO in an insurance company says: “All open systems have their own vulnerabilities. We’re constantly dealing with thousands of vulnerabilities—and identifying them. If we can pull into a single operating environment, we can minimize the effort to keep track of and address security risks.”
This is a strong argument not just for drawing open systems applications onto the mainframe, but also for consolidating multiple, smaller (sometimes legacy) mainframe systems into one large box. The more centralized the IT resources, the easier it is to eradicate areas of vulnerability and deploy consistent policies and processes across the whole IT environment. Other users in the research reflected this sentiment, pointing out that the System z is the only platform scalable enough to support both their consolidation plans and their very large transaction processing or data-intensive scientific applications.
But however strong the logical argument is for centralizing and consolidating, the reality is that this approach is far more accessible to larger users above 10,000 MIPS than small and mid-size mainframe users, whose growth and development are often severely limited by legacy systems that perform well but prove resilient to integration or transformation. As another CIO in the Forrester study says, “We have a large number of legacy applications that we cannot move today, and we have been stable in our mainframe usage.” Such organizations are often forced to encapsulate their legacy resources and face considerably higher software costs than their larger counterparts, since System z software pricing can be very attractive for users who are experiencing MIPS expansion and equally punitive for those with zero or negative growth.
So, the idea of a large, high-growth, consolidated mainframe environment with consistent system management end to end is very attractive, but it may be available only to the “fast-track” organizations. To extend these advantages to the slow-growth legacy mainframers is a more arduous task and requires the concerted efforts of the whole System z community.
What’s Your Take on zPrime?
If nothing else, zPrime from NEON Enterprise Software has generated plenty of publicity and fueled several online debates (check out the fascinating discussion on Bob Thomas’ MainframeZone group on LinkedIn). zPrime effectively allows users to offload “traditional” CP-based applications onto System z Application Assist Processors (zAAPs) and System z Integrated Information Processors (zIIPs) with substantial anticipated software savings. The product itself is technically intriguing, but whether the savings will be realized depends on how IBM and other software vendors react to this unexpected development in IBM’s planned specialty processor strategy. IBM’s System z vice president, Mark Anzani, has made it clear in one publicized letter that, if customers use specialty processors for unauthorized workloads, they must expect pricing associated with full CP capacity. On the other hand, some users are already setting up trials with NEON and examining the cost benefits that might otherwise ensue.
At the very least, zPrime serves to remind us that “new” Java and Linux workloads are effectively subsidized on the mainframe by core CP applications, and that the long-term viability of the platform relies on a balance being maintained between the huge revenue stream that IBM derives from its legacy licenses and the more “competitive” pricing demanded for newer technologies. To what extent NEON will be able to disrupt this balance remains to be seen.