Big Iron Bucks a Trend

Far from transitioning away from Big Iron, customers seem to be doubling down

After a decade of discontent -- i.e., most of the 1990s -- the mainframe came full circle in the 2000s. Even in the midst of what's shaping up to be the worst economic crisis in 80 years, many industry watchers foresee a thriving future for Big Iron. There are, after all, pointers toward optimism.

Take mainframe capacity, for example. Compared to today, Big Iron shops a decade ago were hemorrhaging capacity. In 2000, roughly 3.5 million MIPS were installed on mainframe systems, according to CA Inc.; CA officials expect that customers will buy about the same amount of capacity this year alone. After a September of turmoil in the financial services sector, IBM Corp. officials also predicted a strong 2009 for Big Iron.

Enterprise server sales dropped in both Gartner Inc.'s and IDC's Q4 2008 market tallies, but Big Iron managed to buck that prevailing trend: according to Gartner, System z was the only one of IBM's four server businesses to post a year-over-year gain in revenues. In fact, Armonk's x86 server fortunes took a particular drubbing in Q4 of 2008, plunging by nearly one-third (30.5 percent, per Gartner's tally). The upshot, industry watchers say, is that far from transitioning away from Big Iron, customers seem to be doubling down.

That's industry veteran Joe Clabby's take. Clabby, a former Gartner Inc. analyst who now heads up his own enterprise consulting practice (Clabby Analytics), sees a lot trending in Big Iron's favor. For starters, Clabby says, C-level executives "are striving to find ways to reduce operating costs in these economic tough times, and they are coming to realize that mainframes are not 'old technology' -- but instead are highly efficient platforms designed to run generalized workloads in an extremely secure, high quality of service … manner."

There's another wrinkle here, Clabby continues: mainframe boosters don't have to work as hard to sell Big Iron to their IT chiefs. "IT executives also realize that mainframes are extremely strong consolidation servers -- and they help reduce operational costs related to people-based management, as well as power/cooling costs."

He even goes so far as to use the dreaded "R" word -- i.e., "renaissance" -- in talking up System z's fortunes. "[A] renaissance is now taking place in using the mainframe as a means to drive down IT operational costs and free up money to either balance the books or find new ways to innovate," Clabby argues.

Big Iron Linux, in particular, has been a huge success, with -- at this point -- more than 1,000 off-the-shelf applications, he points out: "The kinds of workload/application mix that can be run on mainframes are expanding rapidly. For example, mission critical application workloads that require advanced QoS … are now finding their way to Linux environments on the mainframe."

Enterprises seem to have learned from the mistakes of the past. A decade or more ago, Hewlett-Packard Co., Sun Microsystems Inc., and other vendors repeatedly took aim at the age, inflexibility, and -- above all else -- purported cost of the mainframe, touting the benefits of mainframe replacement programs. Such efforts typically revolved around flagship Unix platforms -- e.g., HP-UX and Solaris -- but in some cases involved Windows or even Linux.

The hoped-for benefits of such programs never panned out in practice, however, Clabby insists. He cites feedback from Chris O'Malley, the executive vice-president in charge of CA's mainframe business unit, who says shops which have completed Big Iron migration efforts tell CA that they just aren't seeing the promised bang for the buck. "CA observed that many clients are canceling the 'mainframe migration' programs they had instituted or were considering. Some of the reasons for this include recognition that mainframes are capable of acting as huge server consolidation environments for Linux applications."

About the Author

Stephen Swoyer is a Nashville, TN-based freelance journalist who writes about technology.