It’s okay to like your mainframe. Really.
I know what the vendors are telling you about server consolidation, about the ways to simplify your infrastructure. But you don’t have to give into them. You can still be innovative without a mainframe migration project under way. I am saying it because
you need to hear someone else say it: you can keep your mainframe.
Psychologists have taught us about the five stages of grief — denial, anger, bargaining, depression and acceptance. The IT industry is reacting to the recession in a similar fashion, but there is a sixth stage involved — I’ll call it evaluation — which is an important precursor to acceptance. This is when CIOs and IT managers everywhere step back and realize they don’t have to create an e-business enterprise from scratch, and that they possess some assets they should hold on to. This kind of attitude, which was at best considered cautious during the go-go 1990s, is now held up as the sign of a fiscally prudent manager. Mainframes are at the heart of this return to basics.
To discuss mainframes is to tell an IBM story, but Big Blue did not have to be the main character. Amdahl, Hitachi and many others once held an important place in the mainframe market, but they have since concentrated on what seemed like the best-bet growth areas, like storage and services. As its rivals drifted away, IBM has reaped the rewards, reporting that its mainframes saw 30 per cent growth in third quarter of last year. This did not escape the notice of revenue-hungry firms like Sun Microsystems, which teamed with Amdahl IT Services late last year to launch a program that would help mainframe customers migrate to its Solaris-based Starcat servers. But that kind of strategy would be appealing only if IBM were treating mainframes like it does its OS/2 operating system: as a dead technology. Instead, IBM has continued to trot out new mainframe features like capacity backup, SSL-based expansion cards and a networking technology that allows high-speed connections between partitions.
The announcement last month that IBM was introducing a Linux-only mainframe was a bold move that reinforced both its commitment to the big box as well as its support of the open-source OS. Linux and mainframes strike some experts as an odd couple, but the platform’s well-known scalability and reliability offer great potential, particularly when it’s backed by a trusted hardware vendor. Red Hat, meanwhile, has moved quickly to create a mainframe-ready version of its popular distribution. This has been good enough to convince international enterprises like Korean Air and Banco Mercantil to make a mainframe move.
It could be too late for Sun, Hitachi or other vendors to catch up to IBM by creating their own mainframes for the 64-bit era. That’s unfortunate, because there are severe downsides to monopolistic environments, even if IBM didn’t create this one. Enterprises have been trying to chisel their way out of proprietary hardware for at least 10 years. Mainframes may remain a compelling choice for handling applications like bank transactions, but that doesn’t mean IT managers shouldn’t have an upgrade path that gives them some alternatives. Without greater competition, however, IBM is unlikely to provide any.
These are long-term issues, of course, and many people are dealing with short-term realities right now. In this kind of climate, mainframes are poised to thrive. Despite all the irresistible force with which new technologies have tried to wipe them out, they have proven themselves the unmovable objects they were destined to be.