zSeries: Big Plans Ahead?
What can we expect from IBM’s zSeries team in 2006? If history is any indication, it could be an eventful year.
What can we expect from IBM Corp.’s zSeries team over the next twelve months? If history is any indication, 2006 could be an eventful year.
In the past, IBM Corp. has followed up a release of one of its major new mainframe platforms with the introduction of a budget- or midrange-oriented system. The “Baby Z” z800 appeared several months after Big Blue introduced the z900 in December 2001, for example, while the z890 shipped less than a year after IBM’s much-anticipated z900 (or T-Rex) mainframe.
So does Big Blue have a Baby z9 in the works, and if so, what kind of Baby z9 does it have in mind? IBM zSeries execs are playing coy.
“I don’t want to preannounce anything and we’re not ready to talk about specifics, but certainly the system z9 we have today is not going to be the last one [of its generation], and we are continually looking at how we can address additional customer needs, including those of smaller mainframe customers,” says Colette Martin, zSeries program director with IBM.
Elsewhere, Martin says, 2006 could bring relief for mainframe organizations that are struggling to justify the (often prohibitive) cost of running native COBOL or Assembly applications in z/OS. Eighteen months ago, for example, IBM introduced its zSeries Application Assist Processor (zAAP), a dedicated J2EE processor engine for zSeries. zAAP significantly reduced the cost of running J2EE workloads on z/OS. As a result, interest in J2EE on zSeries appears to have skyrocketed.
There’s a chance Big Blue is mulling other, application-specific processor applications for zSeries—possibly for existing COBOL applications, Martin says. “We do believe the mainframe provides significant added value over the competition, and so we want to make sure that it is priced appropriately,” she observes. “We are also looking at specific workloads and looking at specific areas where we may do additional price performance improvements, and I think as time goes on you’ll see us make the kind of enhancements similar to what we did for the zAAP for other types of z/OS workloads.”
zAAP is a J2EE-oriented offering, and J2EE—like Big Iron Linux—is the kind of non-traditional workload Big Blue has been at such pains to cultivate for zSeries. This is of little comfort, however, to mainframe technologists struggling to cost-justify the COBOL status quo. Does IBM plan to introduce zAAP-like engines for traditional COBOL applications?
It’s possible, Martin says. “COBOL applications typically are seen in terms of data-serving types of workloads and back-end types of processing, so if we looked at it, we wouldn’t necessarily look at COBOL per se, but we would look at in terms of database connectivity and such,” she indicates.
IBM’s zSeries team aggressively trumpeted its zNextGen initiative throughout 2005 for good reason: Big Blue hopes to groom as many as 20,000 new mainframe pros by the year 2010. Martin claims that IBM is well on its way. “The specific goal we point to was we said we’ll have 20,000 new mainframe pros in the marketplace by the year 2010, and all estimates at this time say that we will definitely achieve that goal,” Martin comments. “It takes some time to build on itself—you have to start small, you have to get some of the colleges on board. But we’re past that early stage now.”
That’s the good news. The bad news, according to some mainframe pros, is that Big Blue’s zNextGen program gives short shrift to their own career concerns—such as the problem of continuing education.
Martin says IBM is currently exploring ways in which to accommodate the retraining needs of existing mainframe pros.
“One of the things we’ve done to sort of address these in-between skills, for people who are in jobs today and who maybe want to get training on [Big Iron] Linux or [J2EE] WebSphere—we’ve had for years a lot of internal education that has been available to IBMers, and one of the things they’ve done recently is to take some of those classes and open them up so that they are available to anybody who wants or needs to be able to take them.”
The problem doesn’t affect mainframe pros of today, Martin argues. The mainframe technologists of the future will require the same resources, after all. It’s an issue to which IBM will remain focused.
“This just one of the ways that we’re looking at trying to address [continuing education and re-training], and I know that the skills team is continually looking at ways to do this,” she notes. “The ultimate goal, obviously, is to get to the point where everybody says [mainframe pros] are the most knowledgeable and skilled IT professionals in the job pool.”
IT Training: IBM Program Cultivates Talent in College Classrooms
Big Iron Pessimism Persists
Stephen Swoyer is a Nashville, TN-based freelance journalist who writes about technology.