Big Iron Pessimism Persists
With so many major events last year in the mainframe arena, why are so many Big Iron pros still pessimistic about the future?
IBM Corp.’s mainframe unit finished 2005 on a high note, with a strong Q4 (thanks in large part to pent-up demand for its new z9 mainframe systems), a new release of its z/OS operating environment, and another round of publicity stunting on the part of its zSeries Academic Initiative, a program the company introduced to promote Big Iron to the Generation Y set.
Why, then, are so many mainframe pros pessimistic about Big Iron’s future? The problem, some old mainframe hands say, is that IBM still hasn’t done much to address Big Iron’s most intractable problems.
The cost of running COBOL, Assembly, and other traditional mainframe workloads is still prohibitively expensive, they argue, even if next-generation workloads such as zLinux and J2EE have helped recast the mainframe as a viable option for non-traditional customers. Furthermore, while IBM has made a lot of noise about training the next generation of mainframe professionals (with its zSeries academic initiative, for example), some current and former mainframe pros say zSeries educational initiatives are missing in action in their neck of the woods. They also say other key mainframe-boosting programs—such as Big Blue’s Partner in Development program for zSeries—are limited in platform applicability and are not as flexible as they could be.
“I am 45 years old and began my career in the third-party hardware repair business in 1980, working on 360 vintage mainframes and I/O. I have been involved in many scenarios where IBM's mainframe mentality has been a real sore point for their customers and those of us outside IBM who would like to engage in the use, exploitation, and business implementation of those technologies,” says veteran mainframe hand Bernie Schauer.
Like other mainframe technologists, Schauer is particularly critical of the still prohibitive cost of mainframe capacity. “The practice of making the mainframe such a closed and expensive arena has hurt [IBM’s] chance of growing that business with new customers,” he argues.
Ken Sharpe, a mainframe systems programmer with a southwestern state government, agrees—although he doesn’t necessarily fault Big Blue for continuing to charge a premium for COBOL capacity. “I can understand IBM charging the traditional processing customers like they always have for their traditional processing,” he says. “The problem is that we are now developing SQL DB2 stored procedures on DB2 for z/OS—i.e., not COBOL. This means that the processing of this code needs to fall into the new technology category. IBM needs to make this competitive if they want to keep this development deployed under z/OS. This means both software and hardware costs.”
As a result, Sharpe says, his organization is actually paying more to develop and deploy new (that is, non-traditional) workloads on its mainframe systems.
“Our major cost increases are occurring because of use of new technology development under z/OS,” he observes.
Sharpe’s organization isn’t just making investments in mainframe technology, either. Very soon, he says, he and his mainframe colleagues expect to be programming on IBM’s biggest mainframe system to date—the z9. “We are this month scheduled to migrate from a 2064-1C3 to a 2094-109 [z9] all because of this DB2 workload. It would be sad to see IBM lose their customer base because of this,” he argues.
The Trouble with Training
IBM this month once again pumped up its zSeries Academic Initiative (aka, zNextGen), in this case trumpeting a university student mainframe contest, which promised to give students a chance to learn about mainframes and win prizes by completing mainframe challenges. Some zNextGeners have already transitioned to full-time mainframe careers, and Big Blue hopes to cultivate tens of thousands of young zSeries pros.
Mainframe pros generally seem encouraged by the zNextGen program—but some express concerns that Big Blue isn’t doing enough to accommodate the training needs of the existing crop of not-quite-grayhaired mainframe technologists.
“I have seen a real media blitz regarding the move by IBM to create avenues of learning for those getting into computing careers by expanding their presence in academic programs due to the graying mainframer population,” says Schauer. “Unfortunately, I live in Minneapolis/St. Paul and explored their offerings and found there was no program available in my area for someone who would be interested. As a matter of fact, I ended up enrolling in a CCNA class in a local community college instead of looking in that direction.”
More to the point, says Sharpe, there are a lot of things that Big Blue could do to make the mainframe—and z/OS, in particular—a more neophyte-friendly platform. The serverpac was a good step in that direction—zVM maintenance is currently installed with two quick commands. [But] what IBM needs to do is streamline z/OS maintenance more,” he concludes. “I would have expected some progress here since serverpac is becoming a rather aged solution and am surprised that we have not seen any progress from IBM for z/OS.”
An Rx for Mainframe Viability
Stephen Swoyer is a Nashville, TN-based freelance journalist who writes about technology.