In-Depth

Big Blue Keeps Big Iron Out Front

If you're a mainframe booster, there have been plenty of reasons to smile lately. But the announcements aren't just for existing customers.

• By Stephen Swoyer
• 11/09/2004

If you're a mainframe booster, there have been plenty of reasons to smile lately. For starters, IBM Corp. has posted four straight quarters of solid zSeries growth, while mainframe ISVs such as BMC Software Corp. are also reporting extremely encouraging mainframe revenue numbers. Combine that with Big Blue’s deluge of mainframe-oriented news from early last month, and you have a recipe for something—let’s call it prominence—that Big Iron hasn’t enjoyed in awhile.

It all started 18 months ago, when IBM announced its z990 (“T-Rex”) mainframe systems. The z990 servers were the biggest and most powerful mainframes IBM had ever developed. What’s more, Big Blue positioned T-Rex as the lynchpin of its nascent On Demand computing vision, and even unveiled a new utility service that made it possible for customers to license mainframe capacity on an as-needed basis. Because of this, the new z990 systems helped reassure many mainframe customers that IBM was committed to Big Iron for the long haul (see http://www.esj.com/Enterprise/article.aspx?EditorialsID=545).

Shortly after unveiling T-Rex, IBM announced a new “Mainframe Charter” to help demonstrate its commitment to Big Iron and provide a framework that organizations could use to plan their investments in mainframe-related technologies. In addition, Big Blue pledged to standardize IFL pricing on Big Iron systems, so regardless of whether a customer had an older G5 server or a brand spanking new z990, the price of an IFL engine was fixed at $125,000.

This year, of course, IBM announced a new midrange mainframe system, the z890, which ships in a smaller initial capacity footprint than even its baby “Z” mainframe, the z800. In tandem with its z890 release, IBM announced the new zSeries Application Assist Processor (zAAP), a specialized processor designed to speed the execution of Java applications—or of any workloads with Java cycles, such as WebSphere Application Server or DB2. And this summer, Big Blue announced a new version 6 release of z/OS and gave a preview of z/OS 7.

Last month, IBM made a spate of Big Iron-related announcements—including new zSeries roadmaps and technology investments—and touted the creation of two new on-demand centers in Beijing and Montpelier, France. In addition, Big Blue trumpeted increased support from ISVs for sub-capacity licensing on zSeries, citing a new sub-capacity licensing program from Big Iron stalwart Computer Associates (CA).

At the time, zSeries marketing manager David Mastrobattista even talked up a zSeries “resurgence” of sorts, citing—among other encouraging signs—z890 sales that were outpacing the record-setting performance of the z800 during its fist year. “The initial shipments were tracking close to double what the initial shipments were for the 800,” he confirmed.

But IBM’s October zSeries deluge was important for other reasons, argues Gordon Haff, a senior analyst with consultancy Illuminata. “What you really saw a month or two ago was that was really the first time that IBM sort of publicly, externally has really made a sort of strong statement that the mainframe is here to stay, and not just for traditional mainframe customers,” Haff says.

The 1990’s were a tough time for many mainframe boosters, Haff says, and even in the midst of the recovery of the last few years, there was concern among some mainframe customers that Big Iron was not an important long-term platform for Big Blue. “As popular as the mainframe has remained with its base the last couple of years, there’s been a certain sense of it still being outside the main flow of IBM’s product lines and strategies, so to the degree that [IBM] can sort of bring the mainframe into the middle of things, I think that can be very attractive for customers,” he comments.

Of course, you can’t fault some Big Iron customers for entertaining doubts about whether the mainframe resurgence of the last 24 months is sustainable. Big Iron has had its ups and downs, but Haff, for his part, believes it’s here to stay. “Everyone thought it was a bubble when the mainframe kind of spiked up a couple of years ago, and there was lot of talk about how it had to do with product cycles and all of that kind of stuff.

"The fact is, I don’t think it’s a bubble. There seems to be this genuine recognition that the mainframe is better than anything else at doing certain things, and those certain things relate to virtualization and constancy under load and basically the ability to run a lot of things together at one time very well,” he says.

One encouraging sign, said Mastrobattista, is that nearly three-quarters of the mainframe capacity IBM ships is now earmarked for new workloads. “Today, 74 percent of the MIPS that we ship out are in support of the newer workloads, things like SAP, Siebel, PeopleSoft, e-business, WebSphere, DB2, business intelligence, [and] data mining applications,” he said. “These are all what we call the newer workload applications, and in the late 1990’s, that number was below 20 percent of the MIPS shipped.”

IBM has also successfully convinced many new ISVs to develop for zSeries, winning over 50 new ISVs (and 150 applications) in 2003, and—as of October—40 ISVs in 2004. What’s more, Computer Associates last month joined BMC Software and other prominent mainframe ISVs in supporting IBM’s sub-capacity licensing program for zSeries. CA announced a new Measured Workload Pricing plan that’s based on quarterly reports generated by IBM’s Sub-Capacity Reporting Tool (SCRT). No matter what it’s called, the result’s the same: Customers pay only for the mainframe capacity they use.

There’s been other encouraging news on the ISV front as well. BMC, for example, recently reported hearty mainframe revenues, including licensing revenues ($92 million) that were 20 percent better than Q1 of fiscal year 2005—and up 152 percent year-over-year. That’s no fluke, either: earlier this year, BMC posted one of its most profitable quarters in history, thanks in part to healthy sales of MainView, which were up 26 percent over Q3 2003, and up 131 percent sequentially over Q2 2004.

So the bubble probably isn’t going to burst, says Illuminata’s Haff—but Big Blue still has its work cut out for it. “The focus here with these mainframe announcements still remains on the existing mainframe customer base, even if it is about perhaps bringing the mainframe new types of application within that base, but not necessarily going after people who have never had a mainframe in their shop before,” he says. “Perhaps over time IBM will make more of an effort in that regard, but that’s a much more difficult sell. It’s difficult to get people to adopt really any new type of platform, much less something that does certainly have the historical baggage and so forth that the mainframe does.”