Programmer Shortage Threatens Mainframe Future

Software Magazine, April, 2001 by Elizabeth U. Harding

OVER THE LAST 20-SOME YEARS, the mainframe has landed on the "endangered species" list every time a wave of new technology has descended on the industry. Yet the mainframe has not only survived, it has remained popular as a proven, robust platform for running very large mission-critical applications.

Despite the mainframe's tenacity, though, there looms a big question mark regarding its future. While new technology hasn't been able to make mainframes obsolete, the shortage of skilled people just might.

Experienced programmers who have maintained mainframes for years are about to retire and exit the workplace, and young people entering the industry have little or no interest in learning Cobol or CICS. In fact, most colleges and universities don't even offer such classes in their curriculum. So who's going to program the mainframes?

When two of IBM's biggest competitors, Amdahl and Hitachi, announced late last year that they were withdrawing from the mainframe market, Cutter Consortium, Arlington, Mass., did a survey to determine the current demand for mainframes. The survey found that more than half of the respondents still have major applications running on mainframes, and a quarter have more than half of their critical applications on the mainframe.

"It's more significant when you look underneath these numbers," says Ken Orr, Cutter Technology Council Fellow. "Companies put a lot of stuff on client/server and into Web applications. But often these are subsidiary applications, and the core still runs on old mainframes and old databases."

The survey asked companies to name the biggest obstacle to supporting the mainframe in their organization. Of the respondents, 86% cited the shortage of knowledgeable programmers.

"Now it's been over 10 years that the best and brightest people haven't been going into mainframe technology," says Orr. "Old technology doesn't necessarily die because the technology doesn't work. It's that people move on."

Short-term Fix

The problem companies face is how to keep the core applications alive, and what to do next.

Cutter's survey showed that 77% percent of respondents wanted to give the problem to someone else and opted for outsourcing maintenance of their mainframe systems.

"This is a short-term fix," says Orr. "Get the problem off my back because I can't hire mainframe people. This doesn't make the problem go away, and it doesn't add new people to the mix. I hear people say that they're going to outsource to India, but bright, young Indians don't want to work on mainframe computers either."

Bob Wyatt, vice president of the e-solution practice at Keane Inc., a services firm headquartered in Boston, says he worries about the shortage of knowledgeable Cobol programmers.

"I see young kids coming out of school looking for their first job," says Wyatt. "Guess what? The need out there is not for Java, C , and XML because the kids don't have enough experience for heavy lifting. Like it or not, they get pushed into the Cobol world.

"What it comes down to is supply and demand. In terms of the number of jobs out there, 60% to 75% of IT money is spent on maintaining legacy systems. The Global 2000 have far too much money invested in legacy to chuck these applications," Wyatt says.

Keane adopted an intern strategy several years ago, bringing in about 100 people at a time and running them through commercial Cobol training. "We show them that this is how they can get started in this business," says Wyatt.

Changing the Technology

Aware of the need for redefining mainframe computing, IBM is addressing the issue in a couple of ways: 1) designing the hardware so that it largely manages itself; and 2) investing in the popular and portable Linux operating system.

"Skills are an issue across the entire industry and, certainly, they haven't escaped the mainframe environment," says Pete McCaffrey, IBM director of enterprise servers. "We changed the underlying mainframe technology to preserve our customers' investment in their applications."

IBM spent $1 billion and worked on the z900 Series and the zOS for about two years. As a result, McCaffrey says, IBM's new zSeries 900 is very much a self-managing environment that reduces and, in many cases, eliminates the need for human operators.

"Linux is very exciting for us," says McCaffrey. "You've got students coming out of universities trained in Linux, and because Linux is portable, they can develop applications with Linux and then decide to move them from one hardware platform to another. One of the biggest advantages with the Linux momentum will be skills availability down the road for the mainframe environment."

Keane's Wyatt says that he is not convinced that Linux will make huge inroads, however. "Linux has fragmented support, and you need to add a dimension of sophistication to make it industrial-strength," he says. "When you put all these cons together, how much do you gain going to Linux? I see it much more as an alternative to Microsoft."

While the mainframe business is still huge, it's hard to ignore the signs that a decrease may lie ahead. "The big problem is not whether people want mainframes to stay around, but whether there is enough market support," says Cutter's Orr.