Mad Dog 21/21: That’ll Teach ‘Em
March 2, 2009 Hesh Wiener
|
It’s difficult to sell an IBM i server to a prospect who worries that the talent required to use the machine will be in short supply, and even more so to sell an AS/400 or System i shop a new Power System that requires fresh skills. Students have X64 machines and so, usually, do their schools. Power? What’s that? IBM is trying to educate the educators by making access to Power technology cheaper for schools and by supporting faculty members, researchers, and students. And it has a big educational shindig coming up in September. As part of its effort to get more brainpower behind the Power Systems i platform (and AIX boxes, too), IBM has been steadily improving its relations with schools across the United States and around the world. Recently, it got a lot of good publicity from a cooperative program with DeVry University, which has dozens of venues. But the IBM Power Systems division has a long way to go before it can boast the kind of solid academic presence IBM has achieved for its System z mainframes. About five years ago, IBM realized that its mainframes, which were once the gold standard of university computing, had just about vanished from academia. Unix, Linux and to a lesser extent Windows systems had become the Big Servers On Campus for undergraduates, grad students, and the influential group of scholars pursuing MBA degrees. Errors of commission and omission had not only enabled other vendors to build a strong presence. In some cases IBM’s errors not only brought about the removal of big iron, but also left the relationship between IBM and institutions of higher education in tatters. But that has changed quite a bit, and IBM has, in just a few years, begun to reverse the consequences of its policy changes that began, coincidentally, in 1982, about the time IBM obtained a dismissal of the 1975 case brought against it by the U.S. Department of Justice. In 1994, IBM achieved another legal victory, this time gaining a sunset (in 1998) of the remaining aspects of a consent decree it signed in 1956 that constrained activities on the part of IBM that were anticompetitive at the time. That more or less finished off competitors that built mainframe hardware. After that, the only compatible mainframe alternatives came from outfits that had emulation software. IBM’s surviving mainframe rival, T3, which ended up as a dealer and support source in the emulator business, maintains that IBM is more inclined to wield its market power than it ever was. The last commercial IBM competitor still standing before T3 was left alone, Platform Solutions, which developed an emulated mainframe that ran on Itanium servers, took an even stronger position, accusing IBM of tying and other violations of antitrust law. IBM settled its differences with PSI by buying the company and hiring the key PSI employees who might otherwise have fought IBM to the bitter end. The PSI employees were apparently given pretty good employment deals, because they all agreed to keep their traps Trappist. But even as IBM was enjoying victory in the courts and the mainframe market, it was losing influence among the innovators whose ideas often blossomed in academic settings. Search technologies and cloud computing were developed to work with Open software environments running on X64 hardware. Mainframes, which were used at CERN when Tim Berners-Lee built his first Web site, had pretty much dropped out of the picture. People have called X86 hardware “industry standard” for years, which says a lot, and with X64 servers dominating the server landscape, this is largely true. As IBM’s proprietary high-end systems lost visibility, so, too, did IBM’s proprietary midrange machines, computer that once dominated the vast market segment between low-end servers and mainframes. When the AS/400 was first announced in the summer of 1988, eclipsing the IBM midrange triumvirate of System/38, System/36 and System/34, prospective customers viewed it as the quintessential IBM machine for enterprises that were not yet big enough to use a mainframe. But the AS/400 and its successors were, soon after outstripping midrange rivals Hewlett-Packard and Digital, soon overshadowed by the kinds of systems that powered the dot-com revolution. The mainframe lost prominence and then even presence on campus while the X86 iron running BSD Unix, Windows, and Linux got asked to all the proms and startup parties. Without the Big MIPS On Campus mainframe serving as IBM’s flagship in academia, the AS/400 and its periodically renamed descendants never got on the dance card. The changes on campus and elsewhere may be irreversible, and if Big Blue wants to regain its prominence as a hardware vendor, it will have to adapt to very challenging conditions. At the heart of its dilemma is an issue that is making it hard for IBM to get its z and i boxes into academic settings, and anywhere else for that matter: the demand for mainframe and proprietary midrange computing power may be perpetually on the rise, but unit volumes are falling or at best rising very slowly. It is death by Moore’s Law for any machine that doesn’t use high-volume components. IBM’s Power technology may be distinctive, but Big Blue still does not enjoy a reputation for invention, however unjust that might be. Even relatively tiny Sun Microsystems has superior mindshare in academic and research circles, and while Sun deserves enormous credit for the advances it has made in processor architecture, influential developers usually begrudge IBM the appreciation its admirers say it deserves. The only explanation that comes to mind is that IBM is getting punished for its rough treatment of competitors. Sun, in stark contrast, has generally tried to embrace or at least preserve cordial relations with innovators in the computing community. IBM seems to have realized that it has to renovate its hardware business and rebuild its reputation in academia if it does not want to become a vendor of systems that run Linux for new applications and serve as a hospice for legacy workloads. To that end, IBM is consolidating all its hardware (except the X86 machines it could buy as well as build and may soon only buy) onto Power family chips. It is working just as hard to get its proprietary mainframe and midrange systems incorporated into the educational programs of colleges and universities around the world. Big Blue has also made substantial progress toward migrating its mainframes to Power technology. There are still differences between z machines and the standard Power computers used in i and p products–the processor instruction sets are different, but the chips share many other components, and the systems have a lot of common components. IBM also charges substantial differences in prices for memory, I/O devices, and other add-ins on mainframes compared to their functional equivalents on more popular Power boxes. But IBM seems prepared to let these differences fade away as its Power and z processors evolve. IBM’s other challenge is to recapture the hearts and minds of students and their instructors and at the same time reinvigorate the software development community that IBM needs to bring new and improved applications to its proprietary hardware platforms. Outsiders underestimate the considerable progress IBM has made getting colleges and universities to reintroduce mainframes to their computer science students. In 2003, IBM had 24 schools in its Academic Initiative program, today it has 537 and the company believes more than 50,000 students are taking courses that cover mainframe hardware, software and operations. (An awful lot of them appear to be in China.) IBM keeps increasing its effort and now seems to be considering whether it should similarly expand the smaller effort it makes to promote i (and p) systems on campus. A potential turning point for IBM is several months away. IBM has planned an event it calls Power Systems Technical University, which will be held in Orlando, Florida, from September 21 though September 25. IBM is still filling in the details of something like 250 technical sessions and so far hasn’t even decided what to charge for the event, but there is no question that this gathering will be a watershed. If it draws a big crowd and the visitors like it, IBM will almost certainly put more effort and more money behind educators supporting the Power platform. Still, it is a bit odd that Big Blue is calling its gathering a university. Real colleges and universities might not appreciate IBM’s apparent imitation of Humpty Dumpty who, in Through the Looking Glass, said, “When I use a word it means just what I choose it to mean–neither more nor less.” IBM could, of course, rename the gathering before it turns it publicity machine on full blast, which means it would most likely repackage its event during the next few weeks or live with the awkward misnomer, but this probably won’t happen. The possibility that IBM’s PR machine might have (and act on) second thoughts seems pretty remote. Still, offputting name or not, the event is important for i customers, and its success could provide a powerful boost the size of the talent pool supporting the IBM midrange platform. IBM’s overall Academic Initiative, the program that includes the Power Systems Technical University, has i and p offerings for three constituencies: schools, students, and customers. For schools, IBM offers discounted servers (whether leased or purchased), free or cheap access to remote computing environments, free training classes for teachers, courseware (computer-based study materials), access to an online database of opportunities for students and, on a case-by-case basis, access to IBM experts. For students, IBM offers material to supplement ordinary education, such as free software development tools, amusements intended to foster a connection between students and IBM in the Second Life online virtual game and some online services aimed at helping students find career opportunities. The most important career service will be a site where students can post their credentials and achievements as well as details of professional certifications. (Second Life has inexplicably not yet become as boring to IBM’s top scientists as the scientists have become to IBM. Irving Wladawsky-Berger, for a while IBM’s most notorious Second Life time-waster, has retired from IBM and teaches at MIT. Circuit City, which IBM had set up in Second Life, seems to have retired, too.) For end users and resellers, IBM’s Academic Initiative offers connections to the student CV repository. IBM also says it will help users and resellers link up with IBM experts to get knowledge and support from the academic community. To give prospective users of its Academic Initiative offerings a taste of the available services, IBM has set up a demo account at http://aquila.sealinc.org:81/HOD/HOD.html. The username and password for this demo are both systest. Visitors can log in via a 5250 interface in i5/OS or with a DEC VT interface to AIX or SUSE Linux. That VT screen is a reminder that IBM missed the Linux boat . . . and that DEC was completely lost at sea. While this offering represents an increasingly valuable contribution to the future of the Power Systems platform, it represents a risk. The same knowledge that helps i and p users find technical talent is also a potential source of threats to IBM’s installed base. This is particularly the case when IBM emphasizes the use of Linux on Power and the availability of Windows co-processors for Power servers. The same know-how that can be used to build and maintain applications for Power-specific environments can also be the basis of exit strategies. This is the year IBM is going to find out which school of thought prevails.
|
