Evangelizing Mainframe
Print Email

Adapting to the ‘System z Expert’ Decline

The mainframe is going away … never. But the experienced folks who have been managing these complex environments are becoming harder to find. And while there are excellent mainframe-based management tools, more and more mainframe environments are being incorporated into the monitoring infrastructure of the distributed world whose technical staff is not mainframe-educated.

This changing of the guard presents both a problem and an opportunity. The problem is as follows: How do we enable the distributed support team to monitor the mainframe as part of their overall monitoring infrastructure without expecting team members to become mainframe experts? The opportunity is that by accomplishing this goal, we can have a better understanding of the performance of our composite applications that traverse both distributed and mainframe tiers and, as a result, present better service to our customers.

In the past, I’ve written about how essential mainframe systems remain. Over the past two decades, there have been many predictions about their demise and epitaphs regarding their obsolescence. And yet, they continue to be running strong. There are still many people today that suggest the extinction of the species is just around the corner. Experience has shown that the mainframe will continue to play a key role into the next century, if not longer. What is true—and there’s no denying this—is the mainframe now plays a supporting role, and no longer holds onto the leading role as was the case in the past. It’s part of a twisting maze of integrations between applications that span a diverse set of distributed application environments.

However, there’s one aspect of the mainframe that’s becoming harder to find—the “System z expert” who knows the ins and outs of how the mainframe operates. While the mainframe is a staple of modern business, new entrants in the technology job market do not see the mainframe as a launching point. The newest group of IT professionals sees a bigger opportunity to cash in by developing the right social media or iPad app. Thus, the inflow of talent with System z experience is almost nonexistent.

At the same time, the existing pool of talent has been and is continuing to reach retirement age. Had it not been for the recent housing and banking collapse, the current pool would be substantially smaller. The financial crises hit many System z experts in the pocketbooks, forcing them to stay in the workforce longer. While this delay in retirement is supplementing the pool of experts, the trend will end. Within a short period of time, businesses will face a shortage of mainframe-related skills.

The result of this gap, which can already be seen in many environments, is that the teams for distributed systems are increasingly responsible for mainframes. A few companies have offered mainframe-specialized training but were not overly successful. As a result, these new support teams know very little about System z technology. So it’s not surprising that the level of support is impacted. In fact, expecting these teams to become System z experts is not practical.

So, up to this point, this discussion may seem a bit grim. But with every change comes a new opportunity. Instead of looking for what can’t be done, we need to look at the new techniques and solutions that can be brought to market to address the needs of these teams. These will result in the better understanding of the performance of the composite applications that traverse both distributed and mainframe tiers. This has the positive impact of better service to customers. Rather than focusing on green-screen applications that isolate mainframe data and expertise, these new solutions will combine the required data to diagnose and predict problems without regard to platform boundaries. They will also go beyond the event-based integration that has been common for centralization and extract the data required as well as applying rules that allow expertise specific to scope. For example, a spike in Web order-processing time caused by a CICS transaction ABEND due to a lack of disk space on a database no longer takes three different teams to analyze. Existing tools are unlikely to go away, and new ones will supplement them for the foreseeable future.

So the outlook is quite promising. By incorporating the best of both worlds, organizations will continue to leverage their mainframe investments and adapt to the decline in System z experts, while not overwhelming those who remain and still providing their customers with the best service possible regardless of the technology used.

Richard Nikula is vice president, Product Development and Support, for Nastel, a provider of application performance monitoring and transaction management solutions.

Posted: 9/25/2012 12:08:49 AM by Richard Nikula

Print Email

Join Now!