To hear IBM, CA, BMC and Compuware talk about it, 2009 is already shaping up to be a banner year for sales of mainframe wares. The favorite buzz phrase is “recommitment,” which is shorthand for a triad of corporate consumer trends that favor hockey-stick growth in the mainframe market.
According to vendors, customers have abandoned whatever plans they may have had in previous years to move off the mainframe. In 2009, consumers with mainframes are making additional investments in this platform and growing it to handle additional workload going forward. Some, though the numbers are fuzzy, are adding new z10s to their infrastructure, often replacing higher-end UNIX and Solaris platforms or consolidating racks and racks of Linux x86 boxes.
In the process, vendors note that customers are buying management tools like crazy, including automated management for communications and data transfer operations, and for systems and network administration. This increased interest in management automation is intended to not only pare down the number of bodies required to operate the mainframe (already a sparse crew in most companies), but also to handle the increasing workload being placed on the highly capable and increasingly affordable System z platforms.
The second trend: Everyone is buying DB2 toolkits. The mainframe is fast becoming the file server of the enterprise, with companies that already have Big Iron deployed moving lots of distributed systems-hosted databases into massive DB2 databases under z/OS and moving their Oracle and SAP application sets into partitions under Linux on System z. This kind of consolidation bodes well for third-party software houses.
The third trend is the increasing consolidation of vendors in the mainframe software space. What this consolidation translates into is new product sales that are higher than ever before for those vendors still in the mainframe market. Interestingly, only two years ago, this market was largely declared to be “flat and declining” by most analysts and pundits. Even CA was doing some handwringing a couple of years back about the wisdom of continuing its mainframe business. That came to a halt when its new CEO, a fellow with deep roots in IBM, instead “re-committed” the company to support the mainframe and cobbled the 300-odd products together into a more focused mainframe business unit. The company has since been rewarded with seven quarters of strong revenue growth— paralleling the growth in mainframe MIPS consumption estimated by IBM to be 30 percent year over year in 2007 and 49 percent year over year in 2008.
Re-commitment and “renaissance” may be fashionable monikers to describe these collective trends, but I see them as something else: common sense. With cost containment being the primary priority in organizations today, the value of the mainframe can’t help but shine through. No amount of marketecture from the distributed computing crowd can dim this light. However, the trick to sustaining this recognition—and the profits of vendors selling wares in this space—is to learn from past mistakes.
Here are my three simplistic conclusions: First, I would caution the mainframe advocates to chill out with their description of distributed computing as “a collective train wreck” that can only be righted by a return to Big Iron. While these statements contain some elements of truth, they smack of hubris and are sure to alienate firms that might be toying with the acquisition of a new, less expensive, System z box to stand up next to their existing x86 investment. Community mindedness and cooperation to solve business problems in an intelligent way should dominate the discussion to avoid making re-commitment look like “retrenchment.”
Second, document the actual performance improvements and cost savings that accrue to re-hosting apps on the mainframe so we never again have to find our way out of the murky depths of erroneous perceptions that clouded the consumer’s view of mainframes for more than a decade. I’m just not seeing the fact-filled success stories that distributed solutions providers routinely churn out around virtual server consolidation, etc. That really needs to happen if recommitment is to translate into relationship.
Third, do your jobs. If two things killed off the enthusiasm for mainframes in the late ’90s, they were 1) vendors charging too much for software and force marching consumers to next-generation platforms whether they were needed or not, and 2) unresponsive mainframe data centers that didn’t focus on meeting the fast-changing requirements of their business masters. Both provided ammunition to distributed vendors for their replacement case.
Welcome 2009! We’re off to a good start in the mainframe world. Let’s keep it that way with some common sense.