|HOME||PUBLIC LIBRARY||ANOTHER PERSPECTIVE||INFOPERSPECTIVES||CONTACT|
Towards the end of May, Arvind Mahankali, a very smart kid from Long Island, won the National Spelling Bee. His winning word was knaidel. Knaidel is the Yiddish-American term for matzoh ball, variously spelled as matza ball, matzah ball or matzo ball. The top speller was the sixth consecutive desi to cop the prize; soon the winning word might be dacoit. Meanwhile, IBM, with its center of IQ mass in South Asia, can contemplate how it should spell out cloud computing . . . and the possibility that its services empire is the wrong kind in the wrong place.
When Lou Gerstner had a vision of IBM as a services company during his tenure, ten to twenty years ago, his grasp of the opportunity was perfect. He saw that customers were spending a little more on software than on hardware and that they were spending even more on tech staff and outside geeks than either kind thing that they might buy from IBM. Gerstner saw that IBM could boost its business by at least 50 percent if it could get its customers to turn over their glass houses to Big Blue. Moreover, Gerstner understood that IBM's grip on most of its business was due more to the prowess of its anti-antitrust lawyers than its engineers or coders.
By the time Gerstner had been brought in to lead IBM out of a dangerous wilderness it was clear to anyone who cared to look (and to any IBMer who dared to look) that the I in IBM wasn't for innovation. The action in computing had moved from the glass house to the desktop. On the desktop, unlike inside the glass house, IBM had influence rather than control, and not enough influence to exercise pricing power. Gerstner thought, incorrectly it turned out, that IBM could gain a position of strength if it provided strategic software for its not-particularly-differentiated hardware. Even better, Gerstner incorrectly believed, would be a world in which IBM PC software not only ran great on IBM PCs but also achieved killer app status on all the machines derived from IBM's original PC design.
Between IBM's PCs and the Lotus apps IBM bought after prayers at St. Patrick's, a conceit for which Big Blue was ultimately punished, was the stuff that IBM didn't make right, the operating system. It took a long time for IBM to accept how much smarter Bill Gates' outfit was than Lou Gerstner's, plus an insightful but hurtful book or two. But in the end IBM let Lotus grow where it could and die off elsewhere. In the meantime, IBM's opportunity in services turned out to be far larger than the resources it consumed floundering in the PC world.
For a while it looked like IBM had a shot a leadership in the enterprise email business with the server and client software that it offered from its Lotus division. But IBM could never beat Microsoft in that game any more than it could whip Windows with OS/2.
What IBM seemed to have trouble fully digesting is that the thing Gerstner sensed was IBM's insuperable strength has never been clearly articulated by any of Big Blue's bosses. The company's gift-bestowed by founder Thomas Watson and preserved, somehow, ever since-is that it can grab enterprises by the corporate nose and lead them just about anywhere. But it can only do this perfectly when it confines its efforts to the world of beancounting. PCs, and even more smartphones and computing tablets, are not so much about beancounting (although they can be great at it) but about communications in a broad sense that includes non-textual media along with numbers and words.
What IBM missed and to an astonishing extent still misses is that the PC as embodied by Apple and cultural heirs was largely based on the culture of Xerox, not IBM or other companies spawned during the electromechanical record processing era. Xerox completely re-envisioned printing Gutenberg style.
Gutenberg lifted publishing out of the world of scribes by developing systems based on movable type. Xerox abstracted type and engraving into images. Moreover, although it started out with strictly analogue technology, Xerox instinctively understood that analogue xerography would give way to digital xerography as surely as the analog networks in its first PARC offices gave way to packet-switched technology, to Ethernet.
So what Gerstner did to save IBM from the effects of its corporate senility was to set a direction for the company that brought it fabulous success in computer services . . . but at the same time confined it to markets based on dancing with the stars of the boardroom. During the past few years it has been apparent that IBM can't keep growing in services the way it has. It has soaked up just about all the deals there are except for business taken by its rivals. IBM's rivals, like IBM, are heavily invested in the formidable brainpower culture of India, where the brightest few percent of a vast population is thrilled by opportunities in computing. IBM generally has the pick of India's excellent crop because, even more than indigenous rivals like Tata and Infosys, a career with Big Blue still has a cachet. For Indian IT engineers keen on a life in beancounting, the IIT to IBM route is a superior choice.
But you don't have to visit India to notice that the desis rising to the top at the most modern technology companies are usually working in the USA or the UK or Germany, not in India. That might change, but IBM can't wait for the great South Asian Technology Culture Revolution if it wants to feed its top brass three squares in the Armonk cafeteria.
If Ginni Rometty wants to go down in IBM history as a builder rather than a preserver, she will have to take her company in some new directions, which means she will have to take risks. She will have to manage these adventures carefully, because some might pan out better than others. As Gerstner became a hero with his beancounter's concept of services and largely avoided condemnation for his ill-fated acquisition of Lotus, Rometty will have to take a few shots. However she does it, she will have to succeed more than fail. She will not have to get everything right.
Rometty's choices are a little like spelling. Even Wikipedia, which, a true encyclopedia, likes to boil things down, seems unsure about the spelling of matzo, or, for that matter, knaidel (or kneidel). Rometty will have to make brave talk and exude confidence, but she would be fool to think all her guesses will turn out to be correct.
Her first really significant gambit is the acquisition of SoftLayer, now underway. IBM's investment of what various observers reckon is about two billion dollars in a very large privately held X86 hosting company puts IBM into cloud computing in the way most of the world sees it. Notably, until now IBM's operating definition of cloud computing was basically this: Pay Big Blue to use its glass houses instead of your own. IBM might be able to prosper running users' work in its glass houses, but its approach is one of the many zombies that haunt the company, this one the spawn of Service Bureau Corporation, which has been killed many times but always comes back to walk the halls of Armonk.
SoftLayer runs a bunch of server farms and rents the machinery in the form or real or virtual servers to all takers. They are mainly corporate customers who need first-rate service and are willing to pay prices that might be near the high end of a fiercely competitive market for the machines.
SoftLayer publishes list prices for basic servers and additional hardware like extra memory and disk. The prices seem affordable, but even so we suspect big buyers can negotiate a little just so they feel special.
SoftLayer also posts menus showing the main software choices it gives customers. SoftLayer's customers can, for example, pick a one- two- or four-chip machine running any OS from a very substantial list. The company's current offerings include current Windows 2012, mature Windows 2008 R2 or legacy Windows 2003. There's also FreeBSD in 32- or 64-bit, and a ton of Linux distros including RedHat, Ubuntu, Debian, CentOS and Vyatta (for building nework management systems). But wait. There's more. SoftLayer lets customers pop in desktop virtualization from Citrix, plus system virtualization from VMWare and abstractions from Parallels. The software all has list prices in the form of monthly fees. There are also posted prices for things that end up on the shopping lists of most cloud users such as IP blocks of various sizes, many different backbone bandwidths, and other popular items. Prospects who might be moving from internal to cloud computing can quickly see what it costs to approximate (or perhaps precisely duplicate) X86 servers they have inhouse.
Compared to the loosey goosey pricing data services prospective customers are used to getting from IBM, dealing with SoftLayer is crisp and pleasant. Moreover, for an IBM that now really wants to get into cloud services the way it exists in the real world, as compared to the Blue Space, SoftLayer promises to end the boring negotiations over the cost of computing capacity even if they might mean the end of some nice lunches on IBM's expense account.
Now there's no guaranty that IBM will succeed in what is bound to be an unfamiliar and uncomfortable world. But the market for servers in the cloud is, these days, just about the only place computing is big big big and still growing. It's the future, for IBM as it is for every other live player in IT. It's not the world of some Ph. D. character from IBM France trying to drive a Merc on the left in Bangalore without bending fenders. It a world in which a kid with parents from India gets on the path to the university of his choice by spelling Yiddish words better than anyone else.
Ginni, it's gonna be fun. And fun for us to watch you at it, too.
— Hesh Wiener June 2013