Mainframe saga

Three cheers for the engineering team that dubbed IBM’s latest mainframe the T-Rex. The nickname for the new eServer zSeries 990 is an in-your-face jab at critics — including Big Blue’s competitors — who had declared mainframes obsolete, out of date, and headed for extinction.

The nickname has sparked predictable jokes in reply from companies such as Sun Microsystems and Hewlett-Packard, which would like to take away some of IBM’s highly profitable mainframe business.

Yet as Ephraim Schwartz points out in Reality Check, for certain purposes, big iron is still hard to beat.

With the z990, for example, you can pack the equivalent of 1,000 or more Linux servers into a single refrigerator-size box. Compare that to the space, power, and administration required to run an equivalent number of 1U units.  The hardware’s MBTF (mean time between failures) is rated at 50 years. And because Linux is known for stability, IBM says the combination of mainframe heft plus open source OS makes the T-Rex virtually unstoppable.

No wonder IBM edged out HP last year to regain the top market share for computers costing $250,000 and more, according to our sister company IDC.

Predicting the mainframe’s — or any technology’s — demise is a gamble. The Wall Street Journal, reporting on the z990 last month, noted wryly it has now been 12 years since editor-turned-pundit-turned-investor Stewart Alsop predicted in InfoWorld that mainframes would be extinct within five years.

Alsop, now a general partner at New Enterprise Associates, a Menlo Park, Calif.-based venture-capital company, took a lot of heat for that comment at the time and thereafter. He also had the grace to write, when the five-year deadline passed with no extinction in sight, that his prediction had been “completely wrong” and that there was still a place for centralized computing in an otherwise distributed world.

To my mind, both Alsop’s ill-fated forecast and the mainframe-as-dinosaur jokes miss a critical point. Rarely does evolution throw away older forms, even though they may fade in importance. Higher organisms continue to use genes invented by bacteria. The literate and (occasionally) logical human brain still responds to a dizzying number of primitive drives. And although Tyrannosaurus rex no longer roams the prairies, some of its relatives were probably the ancestors of modern birds.

Similarly, it’s easy to imagine the mainframe continuing to play an important role in this industry for the indefinite future — as long as there are advantages to concentrating a lot of computing power under one hood. So here’s to T-Rex. May it stomp its critics and then take flight.