Recently, I read a fascinating article titled “Automated to Death” by Robert N. Charette on the spectrum.ieee.org Website. Charette discusses freak accidents and incidents that have occurred on land, sea, and in the air as a result of (or despite) ultra-high levels of automation. He points out that we’re becoming more reliant on automation on our ships and planes to the extent that when things do go wrong—for example, because of an unprecedented chain of events or exceptional weather conditions—the experts with the right technical expertise are no longer in a position to step in and take corrective action.
Automated systems are only as good as those who design them, and all are prone to failure at some stage. In the engineering world, automation is called upon first and foremost to make machines safer and more efficient but, as Charette argues, there need to be checks and balances in place to ensure specialists can intervene when necessary.
In the business world, automated technology is more often seen as a way of reducing costs and headcount, and one quote in the article sounds alarm bells for those in the mainframe world: “The assumption is that automation is not only going to make [what you are doing] safer but that it will make it more efficient. This creates a rather nasty feedback loop, which means that when adverse events become relatively rare, it is taken as an opportunity to de-skill the people you’re employing or to reduce their number in order to reduce a cost.”
Thankfully, mainframes aren’t often required to avert plane crashes, but they are highly sophisticated and automated, and they support the most business-critical applications running in the world today. Adverse events, of the kind likely to create a major system outage, are very rare in the System z world, certainly compared to distributed server environments. And that’s part of the problem. Mainframe skills in the data center are diminishing not just because many mainframers are reaching retirement age (a much overstated problem) but because recession-stretched financial directors in organizations using mainframes see the systems as black boxes that need little or no technical support—that is until something goes wrong—and are cutting back accordingly.
One recent high-profile mainframe outage in the U.K. was reportedly attributable to staff cutbacks, as a result of which microcode fixes hadn’t been correctly applied. Whether or not this was actually the root cause, it remains the case that when problems occurred, nobody was available to provide the necessary expertise. This wasn’t an isolated incident, and the issue seems to be particularly acute among some of the leading outsourcing companies, which have certainly been keen to reduce staff numbers.
Mainframes can cope with highly complex workload requirements and exhibit unparalleled resilience, but they aren’t perfect. They do need specialized support, but unexpected events will always require human intervention despite the most sophisticated automation. z/OS systems programmers and technical specialists have a highly evolved skillset that’s unique to the System z platform, and companies that believe they can do without them will soon find themselves making headlines—for all the wrong reasons!
Signs of Recovery in the Mainframe Environment?
The 2010 edition of the Arcati Mainframe Yearbook suggests we’re seeing some tentative signs of economic recovery. Seventy-seven of the users who responded to the annual survey said they were expecting some MIPS growth in the year ahead, and overall the level of growth is up from last year’s figures. As in past years, we found that the largest users anticipated a far greater degree of growth than their smaller counterparts, and a worryingly high proportion of sites below the 10,000 MIPS threshold (where commitment to the platform tends to be lower) expected to see capacity growth of 10 percent or less.
Overall, the future is looking positive, with an increasing number of users integrating their mainframe applications with Web services and service-oriented architectures. Approximately two-thirds of sites surveyed felt that specialty processors such as the zIIP, zAAP, and IFL made mainframes more attractive, and 56 percent actually had them installed and were making use of the obvious benefits associated with them. For the full survey results, visit www.arcati.com.
Around the Vendors
CA announced that three of its leading mainframe security management software products—CA ACF2 r14, CA Top Secret r14, and CA Compliance Manager for z/OS r1—are officially in evaluation for EAL4+ certification under the Common Criteria International security standard. Common Criteria evaluation of security products is mandated for commercial information security products purchased by the U.S. government for use in national security systems.
Compuware Corp. unveiled a far-reaching Mainframe Cost Savings Program. The scheme will allow customers to receive a free cost-savings assessment, a roadmap to future mainframe operational costs savings, and even the opportunity to win $1 million!
IBM announced two new enterprise Linux servers, in Enterprise and Business Class variants, that provide attractive, off-the-shelf pricing and configurations for large-scale data center consolidation on Linux. Also announced were two additions to the System z Solution Edition series for Enterprise Linux and Chordiant.