In-Depth
Mainframes and IT: The Year in Review
This year saw more ups than downs, at least for mainframe boosters.
It's become something of a bi-annual tradition: every couple of years, give or take a few months, IBM releases a new version of its System z mainframe platform. Wwithout fail, this latest and greatest mainframe system ups the ante on its predecessor, boasting improved performance, enhanced availability, stronger security, bigger MIPS bang-for-the-buck, and a millions of service unit (MSU) technology dividend that -- for more than a decade now -- has been a can't-miss feature. It's as true in 2012 as it was in 2002: organizations that upgrade to one of Big Blue's newer systems can run their existing z/OS applications on upgraded hardware and actually save money.
That's been IBM's sales pitch since the dawn of the System z age, but it became a little less straightforward late this year following IBM's disclosure -- in late-October -- of new Flat Workload Licence Charges (FWLC). More on that later. .
What isn't to like about Big Blue's newest brainy and brawny mainframe, the zEnterprise EC12?
True, it doesn't far outstrip its predecessor's specs, at least on the basis of raw clockspeed or processor complement (where it improves -- or will improve, given IBM's projected clockspeed ramp-up -- by 15 percent and five percent, respectively), but it does boast substantial cache improvements (one-third more L2 cache; double the L3 and L4 caches of the z196), as well as a new solid state disk (SSD) memory tier that IBM says can help improve its overall availability story. Add in a new Crypto Express 4S cryptographic card, an array of Java-specific enhancements, and improvements to Big Blue's Unified Resource Manager (URM) and you have a strong set of new features. The EC12 even shipped with a Signature First: zAware, an embedded predictive analytic firmware stack that monitors z/OS OPERLOG messages to assess system health.
"This is a stunningly powerful machine, especially coming just 25 months after the z196 introduction," wrote veteran mainframe watcher Alan Radding, on his Dancing Dinosaur blog.
"The zEC12 is intended for optimized corporate data serving. Its 101 configurable cores deliver a performance boost for all workloads. The zEC12 also comes with the usual array of assist processors, which are just configurable cores with the assist personality loaded on. Since they are EC12 cores, they bring a 20 percent MIPS price/performance boost."
Yet Another Big Iron Bust
Recently, we've seen the emergence of another, less-loved bi-annual tradition: the predictable plunging of Big Iron revenues in the months prior to the release of a new System z mainframe. It happened, to disastrous effect, in 2010, just ahead of Big Blue's behemoth-ical zEnterprise mainframe refresh. It happened again this year: System z sales were off in Q4 of 2011, off (by a whopping 20 percent) in Q1 of 2012, and down once more in Q2.
It's the reason IBM moved up its release of the zEnterprise 196 in 2010 -- and it also helps to explain why Big Blue delivered the zEC12 with such breathtaking rapidity this time around.
As Dancing Dinosaur's Radding notes, IBM "broke with its ... historic three-year release cycle to deliver ... [the new EC12] just two years after the z196 first introduced hybrid computing."
This is a predictable phenomenon; what's ironic about it is that the same factors that hich contribute to System z's upgrade value -- viz., improved performance, enhanced availability, stronger security, bigger MIPS bang-for-the-buck, and a millions of service units (MSUs) technology dividend -- likewise contribute to depressed demand for new kit late in a mainframe's lifecycle. Why pay more for less when you can pay less for more?
To avoid the reductio ad absurdum of ever-shrinking release windows, IBM must figure out how to address the problem in a sustainable way.
Speaking of paying less for more, or more for less, Radding reckons that IBM's revised FWLC plan will drive up costs by about 10 percent next year; users idling on older mainframe hardware will likely be hardest hit.
"Since FWLC was introduced IBM brought out five generations of new System z servers. Over that time period the capacity of the largest server available has grown to more than 20x greater than the original IBM zSeries. These generation-to-generation capacity increases have enabled the FWLC software metric to deliver significant price performance benefit over time for customers who upgraded to larger systems," he writes, noting that most FWLC products "are typically older products running on older machines."
These users don't need the capacity -- to say nothing of the bells and whistles -- of IBM's newest mainframe platforms, Radding continues. In this respect, he suggests, "the price increase is IBM's not so subtle way of saying it is time for you to consider upgrading."
Bungling Behemoths
IBM's introduction of its PureSystems hybrid computing platform doubtless raised some eyebrows among some in the mainframe space. At launch, PureSystems supported an array of proprietary and open operating environments, including AIX, System i, Linux, and Windows; it didn't, however, support z/OS or any other IBM mainframe operating environment, for that matter. There was symmetry here, however: the zEnteprise hybrid computing model didn't support System i, after all. In other words, neither of Big Blue's hybrid computing platforms is (as regards support for all of IBM's internal platforms) truly hybrid.
This is a confusing arrangement, and one that IBM must work to address. It's by no means a disaster -- and it's a far cry from the confusion, from the setbacks and blunders, that befell two of IBM's biggest competitors this year. For computing bellwethers Hewlett-Packard Co. (HP) and Microsoft Corp., 2012 was a year of missed -- or blundered -- opportunities.
In fact, for HP it's been two straight years of blundering and missing opportunities -- 2011 was supposed to be the year HP straightened up and worked on flying right.
In 2010, you'll remember, HP tapped former SAP co-CEO Leo Apotheker to fill the void created by the untimely departure of former CEO Mark Hurd. (Hurd stepped down after a female contractor alleged that he had sexually harrassed her; Hurd was cleared of harrassment, but HP investigators determined that he had falsified expense accounts to conceal the relationship.) In retrospect, the Apotheker Era will be remembered as nothing less than a disaster for HP: during the 11-plus months of his tenure, HP's stock price sank by a full 40 percent, and Apotheker's signature or tenure-defining moment -- Hewlett-Packard's $11.7 billion acquisition of UK-based machine analytics specialist Autonomy -- is fast on its way to becoming case-study grist for both the MBA and business law textbook mills.
Autonomy, in fact, is the major reason HP's 2012 is shaping up to be its most disastrous to date: in November, CEO Meg Whitman, who HP's board tapped in October of 2011 to right the company's enterprise business, claimed that Autonomy's former management team had fraudulently inflated its value. According to Whitman, HP might have overpaid for Autonomy by as much as $7 to $8 billion.
Will 2013, then, be the year in which HP finally rights itself?
Microsoft, too, faces an uncertain future. With the conventional desktop computer widely perceived to be at a crossroads, if not at death's door, the translation of Microsoft's Windows franchise to the tablet or mobile context is viewed as critical for its long-term survival.
In October, Microsoft delivered a Windows 8 operating system that met with mixed reviews, with many authorities panning its use in conventional -- i.e., desktop -- contexts, while others questioned its potential for uptake in the enterprise, at least as a desktop OS.
Concomitant with Windows 8, Microsoft kicked off a risky (and unprecedented) foray into client hardware, announcing "Surface RT," its line of branded ARM-powered tablet devices. Surface, like Windows 8, met with mixed reviews; more concerning, however, was its less-than-juggernaut-ical uptake, a phenomenon that Steve Ballmer, Microsoft's characteristically hyperbolic CEO, acknowledged, describing Surface sales as "modest."
One possible culprit for Surface's poor uptake -- other than its $499 price tag and limited slate of applications -- is a good, old-fashioned sales channel bottleneck. Here, too, the story is probably somewhat mixed; poor Surface sales aren't necessarily a function of a shortage of available units, but of Surface's limited retail presence: until early December, would-be buyers could only touch, smudge, or swipe a Surface tablet in one of Microsoft's few retail stores. (Redmond this summer announced plans to open 75 retail stores over the next 2-3 years; by the end of next year, it expects to have just under four dozen retail stores in operation.)
A lot of Surface pessimism has focused on its prospects in the consumer market. It's likely, however, that Surface -- or Windows 8 on tablets -- could fare much better in the enterprise.
BYOD Reconsidered
The putative case for Windows 8, running on Surface RT or on tablets from third-party manufacturers, is also the putative case against BYOD, or bring-your-own-device.
If BYOD has traction in the enterprise -- and it does -- it's because, especially as regards tablets, there's a limited pool of usable, feature-rich, functional tablet alternatives.
To the extent that Windows 8 tablets are able to fill this void, the power of BYOT -- bring-your-own-tablet -- could be diminished accordingly. This is what many who've pounced on Surface RT's "modest" performance aren't getting.
Take tablet market leader Apple Inc., for example. By most standard measures, Apple doesn't do a particularly good job of servicing enterprise iPad customers, at least regarding amenities such as support for automated provisioning and management; in-house, enterprise-scheduled software patch management; in-house enterprise application development; and, of course, access control and security. With the exception of device provisioning and management, these are all areas in which Microsoft has demonstrated competency, if not excellence. At this stage, in fact, Apple's enterprise support ecosystem for its iOS mobile devices is arguably inferior to Microsoft's comparable infrastructure for Windows -- circa 1997.
Fifteen years ago, you'll recall, Redmond's love-it-or-hate-it Systems Management Server (SMS) was its go-to tool for Windows management. Although SMS had its boosters, many enterprises tapped software from Microsoft's arch-competitors -- namely, IBM Corp. and Novell Inc. -- to manage their Windows environments. So it is with Apple, whose iOS platform is still comparatively immature and whose Apple Configurator software is buttressed (and in some cases outstripped) by third-party offerings from AirWatch LLC, BoxTone, Good Technology Inc., Mformation Software Technologies Inc., and ZenPrise, among others.
It took Microsoft a long time to deliver a robust management infrastructure for Windows. Starting with the delivery of Active Directory in Windows 2000, the Redmond firm focused effectively on Windows administration, building a slew of manageability features into its platforms and delivering tools for support and administration. Whatever the initial shortcomings of Surface RT, Microsoft's strategy for Windows 8 tablets is sound: it's betting on its creditable manageability infrastructure; its thriving enterprise software development ecosystem; and its enormous installed base of products, services, skills, and certifications to carry the enterprise.
Whether this strategy pans out remains to be seen.