Power Shift
As software heads for the Web, can Wintel maintain its dominance?
By DAVID P. HAMILTON And DAVID BANK
Wall Street Journal Interactive Edition
The advent of the personal computer in the early '80s turned the technology industry on its head, permanently undermining the mainframe-based supremacy of International Business Machines Corp. and lifting relative unknowns Microsoft Corp. and Intel Corp. to prominence.
Now the Internet and the rise of Web-based applications might just do unto Wintel what it did unto IBM, potentially eroding its PC-based duopoly by shifting the industry's center of gravity to the network itself.
Already, Internet appliances that use no technology from either Intel or Microsoft are proliferating. Sprint Corp. offers a cellular phone that downloads information from the Internet; a handheld device from Palm Computing does much the same. Apple Computer Inc.'s Macintosh computers, long holdouts to the Wintel trend, have staged a resurgence thanks to the popularity of the Internet. And a few manufacturers such as Gateway Inc. and Dell Computer Corp. plan to offer PC-like devices that don't use Microsoft Windows or Intel chips.
All that is the result of an Internet that still largely offers up only static data. Should the trend toward moving actual software applications onto the Internet continue to gather steam, the challenges facing the industry's kingpins could get serious very quickly. For Microsoft, indeed, the shift could fundamentally change its business as thoroughly as any remedy following the recent antitrust ruling against the company.
When applications such as word processing or electronic mail move to the Internet, "people don't have to look at getting a box that runs the equivalent of Windows 98," says Michael Sheridan, vice president of strategic business at Novell Inc. "If it's an Internet application, you'll simply expect it to run."
Wintel, of course, isn't going to simply fade away. Both Intel and Microsoft are taking steps to secure a role in a Web-centered world, although critics doubt they'll be able to dominate the network anywhere nearly as thoroughly as they have the PC.
Intel and Microsoft built their current empire by seizing control of key PC technologies -- specifically, the microprocessor and the operating system. That allowed them to set and enforce compatibility standards for the industry, in essence "locking in" software and hardware developers -- not to mention consumers -- to consecutive generations of Intel processors and Microsoft operating systems.
In addition, by integrating more and more functions into their microprocessors and software while lowering prices over time, Intel and Microsoft brought a significant chunk of PC technology under their sway, allowing them to exercise tremendous influence over the technological evolution of the PC.
Breaking the Lock
What the Internet threatens to do, however, is to break that compatibility lock by making data and, increasingly, software services available to a wide variety of computing devices. In a way, the Internet has become a kind of common interface that allows anyone to access data and services from any computer, whether a Wintel machine, an iMac or a cellular phone.
Some of the starkest challenges to the Wintel model these days are emanating from its longtime nemesis, Sun Microsystems Inc. In the past few months, Sun has made a string of acquisitions and new-product announcements designed to help shift computing away from the PC and toward the Internet. Ultimately, Sun officials hope to turn computing power into a kind of utility, much like electricity or water, that can be delivered to anyone anytime without the need for specific hardware or software.
In August, for instance, Sun announced its acquisition of Star Division Corp., a maker of Microsoft-compatible word-processing and spreadsheet software, and announced that it will make those applications available via the Web this autumn. Similarly, Sun fired a shot across Intel's bow last month, unveiling an innovative microprocessor designed specifically to handle high-speed data and communications for hand-held devices and Internet terminals.
"What's really happening is a seismic shift, from thinking of software as a product to thinking of software as a service," says Greg Papadopoulos, chief technology officer for Sun, based in Palo Alto, Calif. "Someday, we will look back at the notion of someone selling software as a sort of quaint and old-fashioned idea."
Sun has for years promoted the idea of such "network computing," and it certainly stands to gain from the trend, because its powerful computer servers stand at the heart of many Internet sites. By contrast, Mr. Papadopoulos argues that Microsoft faces wrenching change as the computer industry's competitive landscape shifts, since it will no longer be able to count on its technological lock-in to hold its customers.
If applications are housed on the Web, then Microsoft will have to "compete with just about everyone on the planet," Mr. Papadopoulos says.
Microsoft officials acknowledge that Internet computing poses a threat, but they argue both that Sun and like-minded companies exaggerate its probable impact and that Microsoft will be prepared when the time comes. This month, for example, the company laid out preliminary plans to begin offering Office, its dominant suite of productivity programs, as a service over the Internet. Equally important to the company's plans are its recently unveiled initiative dubbed Windows DNA 2000, which aims to establish Windows as an essential platform for building and distributing Internet services.
Side by Side
Publicly, few Microsoft officials claim that Windows will dominate the Internet, and instead say they envision a world in which Microsoft operating-system and application software coexists peaceably with that of competitors. "Windows 2000 is our intellectual property, and we will continue to drive forward with that," says Bill Anderson, head of Web application services for Microsoft, based in Redmond, Wash. But, he adds, "in a heterogeneous environment down the road, it will become increasingly difficult to interject proprietary standards in a Web-based world."
So, for instance, Microsoft has embraced an industrywide standard for distributing data known as XML, for Extensible Markup Language. XML seems likely to become the common language of electronic commerce, making it possible for businesses to exchange in a universal format purchase orders, product descriptions and other minutiae important to e-commerce.
Microsoft has driven aggressive efforts to standardize the use of XML across the industry, even establishing a clearinghouse of XML data types called biztalk.org (biztalk.org). Some critics have been surprised at the company's embrace of the standard; many expected Microsoft to attempt to subvert it by adding proprietary extensions that would work only on computers that run Windows.
But so far, the company's approach to XML differs substantially from its defensive reaction a few years ago to Sun's Java technology, an earlier attempt to break Microsoft's lock-in by making it possible to transfer software programs across incompatible computers without modifying them. E-mail disclosed as a result of Microsoft's numerous legal tussles has shown that officials from Bill Gates on down feared Java's threat to Windows; as one Microsoft foe put it, company officials set out a strategy to "embrace, extend and extinguish" Java by building in extensions that would tie it closely to Windows. Microsoft eventually had to abandon that strategy when Sun sued and obtained a ruling that forced Microsoft to hew to Sun's Java standards; the matter remains before the courts.
This time, Mr. Anderson insists that Microsoft will adhere to industrywide XML standards. "It benefits us to be a good player in XML space," he says. Most Microsoft-inspired extensions to the XML standard, he says, will be accepted industrywide; any exceptions will be "one-off" solutions tailored to solve particular problems. "We're saying we're going to take that framework, build on it and extend it, and make sure it's robust for the Windows 2000 platform," Mr. Anderson says.
Microsoft, however, envisions XML as much more than a simple data-description language; instead, it considers the standard a way of letting programs communicate with each other across networks of otherwise incompatible machines. That job currently requires the use of Java or other similar technologies that create a layer of compatible "middleware" that allows programs to communicate. Microsoft competitors such as Sun and IBM consider the company's passion for XML little more than a thinly disguised attack on Java and other middleware technologies.
Indeed, some critics scoff at the notion that Microsoft intends to cooperate with the rest of the industry indefinitely. Fear of the Internet explains "why Microsoft is rushing so quickly to embrace the Web, to extend it in proprietary ways and get people to use those proprietary extensions," says Dan Kusnetzky, an analyst with International Data Corp. "Once you do, you are tied to Microsoft, which is trying to own the Web in a way no other provider is really trying to do."
Another important element of the Windows DNA 2000 strategy is what the company calls "megaservices," Web-based services for companies that operate Web sites. Some, like Hotmail, have already become successful Web businesses in their own right. Others include the MSN Messenger instant-messaging technology, the LinkExchange system for promoting sites, and Windows Update, the basis for a broader service that manages software upgrades remotely. Microsoft has also launched a site, bCentral.com (www.bcentral.com) that offers other services for small businesses, such as opportunities to advertise on the Web or create e-mail newsletters for customers.
The idea is to let customers link Microsoft and non-Microsoft services to create their own broader services. For example, an online shoe retailer might integrate its suppliers' inventory systems, a sales-tax calculator from another Web site, and a Microsoft megaservice called Passport that manages credit-card authorizations and fills in online registration forms. Online customers might never know that their shoe purchase had taken them to half a dozen servers distributed across the Internet.
Microsoft executives say they haven't worked out exactly how to get paid for such megaservices. Many of them won't be directly visible to consumers and so aren't suitable as advertising vehicles.
Still, Microsoft is committed to the concept. "For the longest time, we've been about Windows," says Paul Maritz, group vice president of the Microsoft division targeted at software developers. "Now we're extending that out further into the megaservices world."
The Windows DNA 2000 strategy isn't assured of success, of course. Microsoft has already stumbled with the first step, which is simply bringing its new Windows 2000 operating system to market. That operating system is Microsoft's biggest attempt yet to make inroads in the high-end server market, now dominated by Unix systems from Sun, IBM and Hewlett-Packard Co. Already a year behind schedule, Windows 2000 won't ship to end users until February.
Microsoft potentially faces even bigger problems if popular programs like Microsoft Word migrate to the Web, as Sun hopes. The software giant depends on these and other parts of its Office suite of software for roughly 40% of its revenue, and has a great deal to lose if these products morph into free Web versions, as electronic mail and calendars have already.
But Microsoft is getting ready to counterattack with its own software rental scheme, dubbed Microsoft Office Online, that is expected to be up and running in the first half of 2000. The company's bCentral.com Internet site will offer Office to small businesses directly, and Microsoft will work with other service companies that want to develop their own plans to rent Office. The company hasn't spelled out any definite pricing scheme yet, but its executives say there is no reason why fees for renting out the software can't be set in a way that brings Microsoft at least as much revenue as it gets from Office today. "We don't see this in any substantial way changing our Office business," says Bob Muglia, a Microsoft senior vice president.
Indeed, some Microsoft executives say the Web-based approach could actually extend Office's life. The sales price of Office has declined steadily for more than three years, and more copies are sold as low-priced upgrades than as new installations. And because old versions work well enough for many people's basic computing needs, motivating computer users to upgrade to each new version has become a major challenge.
Microsoft President Steve Ballmer says he would actually prefer to sell Office on a subscription basis, and is in discussions with several Internet service providers to do just that. "We could get $100 to $150 a year from people," he says, "and if they don't pay, they get cut off."
Intel's Strategy
Intel, based in Santa Clara, Calif., faces similar challenges from the Internet. "There's a lot to be paranoid about in that space," says Craig Barrett, Intel's chief executive officer. "We're still very dependent on the PC marketplace for a lot of our revenue and earnings."
Still, Mr. Barrett and other Intel executives argue that they are well-positioned for the coming revolution. For starters, they say, Intel has refined its marketing message. The chip giant, which long prided itself as producing the "brains" for the world's PCs, now calls itself a supplier of "building blocks for the Internet economy."
Behind the marketing shift lies a three-pronged strategy. First, Intel plans to extend its sway over microprocessors to high-end servers, which now mostly use a variety of incompatible chips made by the computer manufacturers themselves. It is investing heavily in chips for networking communications, having laid out more than $6 billion over the past two years to acquire a variety of communications-chip companies. And it is promoting an inexpensive low-powered chip called StrongArm as a contender for set-top boxes and hand-held devices that may serve as Internet connections for millions of people.
While the Internet may also break Intel's technological lock-in, its executives argue that the company can still succeed by driving down chip costs through manufacturing prowess and sheer scale. "There's still a huge market growth potential out there," Mr. Barrett says. Whatever happens with the Internet, "people can't do this without hardware and components, and that's our specialty."
Like Microsoft, Intel has hit a few potholes along the way. Its new high-end microprocessor, recently named the Itanium, is more than a year late and won't be shipping in quantity until the second half of next year. Many analysts expect that the Itanium will be slower than competitors and that it won't be until its successor, now code-named McKinley, is released in mid-2001 that Intel can hope to make much headway in high-end systems. (Intel says it's confident that Itanium will be accepted in the marketplace.)
Similarly, its Pentium line faces renewed competition from Advanced Micro Devices Inc., whose Athlon chip is now in a neck-and-neck race for fastest PC microprocessor with Intel's Pentium. If AMD succeeds at pushing Athlon into the more profitable high end of the microprocessor market, Intel might have to either sacrifice profits in a price war to keep AMD down or accept sharing the field with its rival.
Things look somewhat brighter for Intel in communications chips and at the low end, where the latest version of the StrongArm has received generally positive reviews and is even starting to win customers -- most notably Nokia Corp. of Finland, which has agreed to use the chip in a new digital-television set-top box. But the market for such chips is likely to be far less profitable than that for PC microprocessors. Some Intel officials even doubt that there's likely to be much of a market for such Internet appliances.
Leap of Faith
"History is not encouraging in that regard," says Andrew Grove, Intel's chairman. "The idea of appliances didn't start six months ago -- people have talked about them for years and years. What's reached even 1% of PC penetration? Just two devices, cell phones and the Palm Pilot."
Indeed, some skeptics argue that while the Internet is important, it's unlikely to remake the entire computer industry very soon -- if at all. Even Sun's Mr. Papadopoulos says that for network-centered computing to really take off will require what he calls a "leap of faith" on the part of people who are used to thinking of the PC as the most convenient place to store their personal and work-related data.
That change, he suggests, isn't entirely unprecedented; he likens it to that point earlier this century when ordinary people first decided it was safer to put money in banks instead of under their mattresses.
Others argue that the global computer network remains far too unreliable to provide the kind of seamless utility-style computing Sun and others talk about. Worse, reliance on the network could reduce the independence of computer users, who might end up reliant on network managers who won't necessarily have their best interests at heart. All that suggests to some that the PC is likely to remain the workhorse of the computer industry for some time.
"I'm peeved when my PC crashes, but I'm really bothered if the network is not available," says Gary Herman, director of the Internet systems laboratory at H-P's corporate laboratories. "There are a lot of things network people won't get around to fixing, which means there are many things users will require locally [on a PC] even if it's more useful to do them centrally.
"When you give up control like that, which is what happens when you run applications centrally, it's just not going to happen without good reason."
-- Messrs. Hamilton and Bank are staff reporters in The Wall Street Journal's San Francisco bureau. |
