In-Depth

Linux Driving Mainframe Shipments

Removing limits keeps customers, report says

• By Stephen Swoyer
• 04/10/2003

Writing in a recent research report, John Phelps, vice president and research director of servers and storage with Gartner, cites a number of trends that support the long-term viability of the mainframe, including IBM’s willingness to invest in z/OS and zSeries; mainframe sales to first-time customers; the growth and popularity of Linux on the mainframe; and the persistent growth of large mainframe shops, which continue to add MIPS.

Although Phelps acknowledges that Big Blue has curtailed its investment in the “low-end of the market”—VSE, VM, and systems smaller than 80 MIPS—he speculates that this has occurred to allow continued investment in z/OS and zSeries: “Although total investment in the mainframe is shrinking to match shrinking mainframe revenue, IBM continues to consider the business worthy of investment and has shifted to a smarter, more focused investment model.”

The result, Phelps writes, is that IBM will introduce substantial performance and scalability enhancements to z/OS and to zSeries hardware, including, he predicts, “the addition of 16-way symmetric multiprocessing by year-end 2004 …with plans to start with a 24-way bit architect for a 64-way or larger system.”

In addition, Big Blue will work to remove other limitations, including the current 256 I/O channel limit and the 15 LPAR limit, by the end of 2004.

Phelps notes that mainframe sales took a nosedive in the 1990s—plummeting to about 30 or 35 a year—but says that IBM has moved considerably more hardware in the new millennium. In 2001, for example, Big Blue brought 70 new mainframe customers onboard; in 2002, he writes, “there were another 100 or more new customers.”

At the same time, however, Phelps put the number of defections away from Big Iron in 2002 at “probably more than 200,” suggesting that most of these customers were running at less than 40 MIPS, with many of them in the 3-15 MIPS range. Conversely, most of the new mainframe systems sold by IBM were in the 80 to 1000 MIPS range.

Another factor buoying the prospects of the mainframe is the popularity of Linux, which has actually helped to drive mainframe shipments, accounting for 15 percent of the new MIPS shipped in 2001 and 20 percent in 2002 by Gartner’s estimation. According to Phelps, much of the Linux-driven uptick in 2002 happened in the form of trial systems, however.

All told, Gartner says that more than 200 IBM mainframe customers have deployed at least one Linux application on mainframe systems in production environments. Another 400 are in the process of implementing Linux-on-mainframe applications, or are at least evaluating doing so.

Still another good sign for the mainframe’s long-term viability is the overall health of Big Iron in large enterprise environments. Gartner’s research indicates that enterprise shops with more than 10,000 MIPS continue to grow, and Phelps writes that IBM is wisely encouraging growth by fine-tuning “its software pricing approach to make it cost-effective for those customers to add new and legacy-related applications because of the attractive incremental costs.”

The mainframe is not without its problem areas, which include—Gartner believes—a lack of new applications and diminished enthusiasm among some ISVs. Moreover, writes Phelps, mainframe pricing models—particularly among ISVs that tend to treat the mainframe as a “cash cow”—have hampered the platform’s TCO story: “One reason software costs have become a major issue is because most IBM mainframe software is priced on a capacity-based license and the mainframe runs a consolidated environment.” (For more on capacity-based licensing, see http://www.esj.com/news/article.asp?EditorialsID=408.)