Sun and Oracle Channel Sun How to Buy Log In United States [Change] English

Articles from the Past 25 Years

Table of Contents
 
 

Finance; Briefs

Business Week
December 20, 1982

Summary: Kleiner Perkins plans to start investing the fund promptly, plowing $2 million into two computer equipment companies -- Sun Microsystems Inc. and FCI Magnetics Inc.

Back to top

Sun Unveils 32-Bit Workstation

Computerworld
November 7, 1983
Mountain View, CA

Sun Microsystems, Inc. has announced two 32-bit, engineering-oriented workstation/processors based on the Motorola Corp. 68000 microprocessor. The units operate under the firm's version of the Unix operating system.

Called the Sun-2 family, the units consist of deskside and rack-mountable workstations that offer dedicated computing power plus the ability to use the units in a local-area network, such as Xerox Corp.'s Ethernet.

Aimed at software development, document preparation and computer-aided design and manufacturing applications, the workstations, called Sunstations, feature local-area network communications as a standard capability.

Local-network hardware and software permit several types of resource sharing among clusters of Sunstations, such as common printer service and common modems for remote telecommunications.
The Sun-2/120 can be configured either as a network node or a self-contained, stand-alone system. The Sun-2/170 is a rack-mountable system intended to act as a file server equipped with a disk drive, tape drive or other peripherals, according to the vendor.

Both models use an Intel Corp. Multibus backplane, with nine slots in the Sun-2/120 and 15 slots in the Sun-2/170. A standard Sun-2/120 workstation, including a processor, 1M byte of main memory, a bit-mapped graphics display, keyboard, mouse, Ethernet interface and all software, costs $16,900. A stand-alone configuration substituting a 42M-byte formatted disk drive and a 1/4-in. tape unit for the Ethernet interface costs $16,300, the vendor said.

A standard Sun-2/170 workstation, including the processor, 2M bytes of main memory, Ethernet interface and software, costs $20,900, Sun Microsystems said from 2550 Garcia Ave., Mountain View, Calif. 94043.

Back to top

High Speed Management for the High Tech Age

Fortune
March 5, 1984
Susan Fraker; Andrew Kupfer

HIGHLIGHT:
High-speed management doesn't mean running faster on the old treadmill. It's a briefcase full of techniques for coming up with new products, making sure that they're what the customers want, and getting them to market in time to cash big.

RAPIDLY CHANGING technology, quick market saturation, unexpected competition -- these all make succeeding in business, particularly a high-tech business, harder than ever today. Managing well, in the classical sense, isn't enough. You have to manage differently. The skills that make up the new technique -- call it high-speed management -- aren't easy to master. Business schools don't teach them. But learning them is becoming increasingly imperative, even in industries not commonly regarded as high tech. In this instance, an old phrase has taken on new meaning: he who hesitates may indeed be you know what.

At the heart of the current need for change is the fact that product life cycles are getting shorter. "I can't document it, but every industry we look at seems to be undergoing shorter cycles," says Joel Goldhar, dean of the business school at the Illinois Institute of Technology. "All of a sudden, industrial products are like Hula-Hoops." The microcomputer was almost unheard of seven years ago. The first commercially successful machine could process only eight bits of information at a time -- a bit being a one or a zero in binary code. But no sooner did the eight-bit micro come to market in 1977 than the 16-bit micro was on its way in 1981, followed by the 32-bit in 1983. Eventually the market for micros will mature, and the time between new products will stretch out again. This is exactly what happened with electronic calculators (see chart, page 64). One moment, infancy; the next moment, old age. Blink and you may miss the market.

Dominique Hanssens, a professor at UCLA's Graduate School of Management, has studied the life cycles of electric appliances. Years ago, he says, the market for refrigerators, say, took over 30 years to mature; the market for newer appliances like microwave ovens has taken about ten years.

"Manufacturers face the end of a growth era much sooner than they used to," says Hanssens. Why-- The rapid rate of technological change, plus the easy availability of credit and the power of mass communications and advertising. "In the old days," Hanssens says, "people who bought new products were what we call 'innovators' -- the wealthy and well educated. It took a while for something to diffuse through the population. Today, very soon after a product hits the market, Mr. Everybody has it."

If they're to survive, companies whose products live shortened lives can't do business as usual. Take pricing. The conventional view has been that a business in its growth phase doesn't aggressively compete on price. If sales are booming, why cut prices-- That comes in the mature phase, as demand slacks off.

Of late, however, consumer electronics products have gone through the cycle so fast that a new one becomes universally distributed almost overnight. That's why in businesses that seem to have been born yesterday one sees cutthroat price competition and other characteristics of mature industries -- for example, competitive advertising ("my computer, X, is better than yours, Y"). In the semiconductor business, production costs have declined so radically that it's been difficult to maintain any kind of stable pricing policy. Hanssens says the whole process has become so distorted that he wonders whether the traditional notion of product life cycles will survive the 1980s.

DEALING with these distortions has contorted many a company. Within the last year Digital Equipment, Atari, and Texas Instruments all had large, unexpected losses. Atari's problems provide the clearest example of a company caught off guard. Virtually no one in the industry realized how quickly the market for video games would mature. Atari based earnings projections on the assumption that demand would grow at the same rate as in the past, and that the company would hold on to its share of the market.

How does a company avoid such unpleasant surprises and prosper-- Executives who have survived and surmounted the challenge report that they've learned several lessons.

The first is the necessity to think constantly about new products, and then to back this thinking with investment -- fast. "If you become complacent in this business, you're dead," says Steve Blank, a marketing director of Convergent Technologies Inc., which produces desktop work stations. "It's the next product and the next and the next that keep you alive." In 1981 the consulting firm Booz Allen & Hamilton surveyed 700 U.S. companies about product development. These companies said they expected that nearly one-third of their profits in the 1980s would come from new products, compared with slightly over one-fifth in the late 1970s. The median number of new products introduced by the companies from 1976 to 1981 was only five. The companies project that from 1981 to 1986 they will introduce twice as many.

A good new-product strategy requires a fat R&D budget, and a lot more. It demands continuous close contact with customers, careful monitoring of the competition, wise marketing and manufacturing, and impeccable timing.

"Stay close to the customer" -- the advice comes straight out of the best-seller In Search of Excellence. No company, high tech or low, can afford to ignore it. "Successful companies always ask what the customer needs," says William P. Sommers, a Booz Allen executive vice president. "Even if they have strong technology, they do their marketing homework." Hewlett-Packard, which used to design scientific instruments by asking the guy on the next bench what kind of device he would like, now asks its customers. H-P's personal computer division even ran focus groups -- in which potential customers sit around and talk about their needs and interests -- to see what computer buyers wanted. Focus groups-- At Hewlett-Packard-- "We'd never heard of them," admits Edward McCraken, H-P's executive in charge of business development. "But in using them we discovered overwhelming support for our touch screen." With this device, the computer user can give directions to the machine simply by touching a spot on the screen's face. "The popularity of the touch screen was contrary to what many of us at the company expected," McCracken says. "I grew up typing. I didn't want a touch screen." When its customers spoke, Hewlett-Packard listened. Instead of making touch screens an option on the HP150 as planned, the company made them standard.

Smart firms try to stay almost as close to the competition as to the customer. When products and manufacturing processes change all the time, it's crucial to keep up with the investment strategies and costs of rivals. If Texas Instruments had snooped a little, it might not have taken such a bath in the home computer business. According to a former TI executive, the company mistakenly assumed that its costs of making the 99/4A home computer were in line with the costs of competitors, chiefly Commodore International. "TI thought it could withstand the price pressure because its costs were competitive," says this executive."It was wrong. You don't get into a price battle, particularly one where you bet on capturing enormous volume, with a competitor who has significantly lower costs." TI ended up losing $660 million in the home compuer business before withdrawing last October.

Quick, effective product development requires closer cooperation among design, manufacturing, and marketing than most companies are used to, not to mention an executive with the skill of a Balanchine to choreograph the proceedings. Companies in which the design team continues to toss a product over the wall to manufacturing risk adding years to development. Rolf Brauchler, director of marketing at Quantum Corp., a Silicon Valley producer of Winchester disk drives for computers, says a design team often fails to understand production and comes up with products that are hard to make. Quantum's solution: its design engineers help develop the manufacturing process.

Deere & Co., which has been a leader in exploiting computer-aided design and manufacturing for farm equipment, recently announced its first factory where the manager of product engineering and the manager of manufacturing engineering will be the same person. "In the past these tasks were done serially," says Steven C. Wheelwright, a professor of management at Stanford's Graduate School of Buisness. "It's always easier to see the next step once you finish the last one. But now they must be done in parallel."

Always the key is to move quickly. Hewlett-Packard has every division spend at least 40% of its time on short-term projects. A short-term project used to be anything that could be completed in three years; now it's defined as one that will take a year or less.

Two Japanese electronics manufacturers, Matsushita and Casio, have further refined the science of speedy product development. After determining what their customers want and designing a product, they prohibit major changes. "An engineer can always see improvements to make," says Robert J. Conrads, a partner in the McKinsey & Co. consulting firm. "But it would delay development to incorporate every improvement. The new ideas can be included in the next generation."

Matsushita's strategy for its Panasonic and Quasar brands has been particularly devastating for Sony, which takes a more painstaking approach to R&D. Sony is still a brilliant innovator. But it has had little time to reap the rewards of innovation because rivals follow hard on its tail with better-selling products -- in the videocassette recorder business, for example.

MATSUSHITA'S EXPERIENCE proves an important point: a company need not be first with a new product. For a follower, IBM has done okay with mainframes and micros, though it missed the minicomputer market almost completely. Companies with fewer resources than IBM should try especially hard to be fast followers, however. The advent of short life cycles means that the big money is to be made quickly; a product that arrives six months late could easily lose six months from its payback period. Digital Equipment and Data General, two of the largest manufacturers of minicomputers, have both been troubled by late products.

A related problem affects companies that announce products too early. Osborne Computer Corp. hit the jackpot with the first portable computer. But, predictably, other outfits began introducing competing products. To hang on to customers, Osborne announced its second-generation portable long before the company could ship it. The strategy backfired when dealers stopped buying the first machine to wait for the more advanced version. That created a cash flow crisis, and matters worsened when the new machine turned out to be a disappointment. Ultimately the company went under. Sun Microsystems Inc., a privately owned California manufacturer of engineering work stations, circulates articles about Osborne to its employees, as a lesson in what not to do.

Companies versed in high-speed management would rather cannibalize their own products than have a competitor do it. They take as a given that their position in any market is ephemeral. "I would argue against trying to squeeze every last dollar out of a winner," says Julien Phillips, a partner in the McKinsey firm. "It's more important to figure out what's going to replace that winner." Rather than making incremental changes to upgrade its electronic bank teller terminal, for example, NCR Corp. introduced an entirely new, programmable machine.

STRATEGISTS ARE also coming to realize that two or more heads, and pocketbooks, may be better than one. Many companies now establish joint ventures with other organizations or finance venture capital firms. "These guys aren't looking to make easy money," says Edward Roberts, who directs MIT's program in the management of technology. "They're looking for new ideas." Corning Glass Works has projected the joint venture technique with Dow Corning (making silicon products with Dow Chemical), Genencor (enzyme research with the biotechnology company Genentech), and Siecor (optical fiber cable with Siemens of West Germany), to name a few. Yoshi Tsurumi, a professor of international business at the City University of New York, argues that as competition in almost every high-tech business intensifies, "any company that limits its R&D resources -- money, people, ideas -- to its home base will be restricted."

Acquisitions provide another way of expanding a company's technological ambit. Small companies or entrepreneurs working alone account for a large share of successful innovations. Big companies have more capital, plus well-established marketing, sales, and service channels. The trick is to combine the two sets of strengths successfully. It's tough. Big companies frequently, if inadvertently, smother the small outfits they buy. Zilog (microprocessors), Vydec (graphic terminals), and Qyx (electronic typewriters) all had state-of-the-art products when Exxon Enterprises, a division of Exxon set up to get into high tech, bought them in the 1970s. But Exxon had trouble managing the product development process, and its small ventures never lived up to expectations.

Exxon's experience underlines a point to remember about managing change: it requires accommodations -- sometimes wrenching ones -- by the corporate culture. Innovation entails disorder. Companies that want to legislate disorder out of the process won't succeed. Julien Phillips of McKinsey says banks in particular must make profound changes in the way they do business. "Banks need people who can think about solving a customer's financial problems," he says. "That's as valuable as the money they lend. But that means banks need to hire people who understand a customer's business a lot better than the credit officers do now."

An influx of people with different mind-sets is often disruptive, but every company must reexamine hiring policies as its business changes. As growth takes off, for example, high-tech firms may need to pick up marketing talent from consumer goods companies -- as Apple, Atari, Hewlett-Packard, and Microsoft have. Personal-computer technology is so widely available that marketing will play a major role in determining who succeeds or fails. According to the former executive from Texas Instruments, TI's decision to base its home computer division in lackluster Lubbock, Texas (300 miles from Dallas), made it impossible to attract enough marketing talent, not to mention skilled circuit designers and software people. "They just didn't have the people they needed to stay on top of things'" he says. Scott McNealy, one of the founders of Sun Microsystems, advises growing outfits to hire overqualified individuals. "These people don't just have to do their jobs," he explains, "they have to create the company."

The most difficult challenge may be managing creativity and the people who embody it. Companies that have mastered high-speed management try to keep the mental light bulbs on by establishing small teams to design, manufacture, and market new products. Whatever they're called -- entrepreneurial groups, independent business units, skunk-works -- these teams remove the bureaucratic straitjacket from product development. Robert Conrads of McKinsey has found that separate business units offer one of the best ways of cutting product development time. But he cautions that they represent a major and potentially nervous-making departure from the way things are usually done at large companies.

"It's so easy for management to compromise and say, 'Let us do your financial planning,' or 'Let's share engineers,'" he says. To be successful, independent business units must be small, hardworking, and preferably located somewhere away from the normal corporate premises. Matt Sanders designed Convergent Technologies' newest product, an electronic spreadsheet called Workslate, in a former credit-union building a few miles from headquarters.

Abbott Laboratories and NCR both like the smell of skunkworks, and see in them a way to force decision-making further down in the organization. "I just assume decisions will be made two or three levels too high," says Jack Schuler, a group vice president of Abbott. "So I use the Mao Tse-tung approach. I put the squeeze on middle managers by talking to every new employee to explain that decisions should be made as close to the customer as possible. If the middle managers intervene, the young tigers say, 'Hey, that's not what Jack Schuler said it would be like.'"

The companies that cope best with rapid change try to anticipate it, even if this means reformulating strategy and attempting to alter the corporate culture. For years Monsanto made its money producing commodity chemicals. But in the 1970s oil prices rose and oil companies moved into petrochemicals. "It became apparent," says Executive Vice President Earle Harbison Jr., "that the future of a chemical company in commodities was not sound." Monsanto pulled out of businesses employing $1 billion in capital to move in two other directions: biotechnology and fabricated manufactured products, such as Astroturf.

THE COMPANY devised a clever way of altering its culture to facilitate the change. It began investing in Biogen, Genentech, Cetus, and Genex -- all start-up biotechnology firms. Then it invited scientists from some of these outfits to company headquarters to train Monsanto employees. "Monsanto had an investment in people [from the start-up firms]," says Edward Roberts of MIT. "It softened its inside beachhead by bringing those outside guys in." The next step was to finance biotechnology projects inside the company. Monsanto tripled its basic research budget and is building its own research park near St. Louis. To further expand R&D, Monsanto formed joint ventures with the biotechnology firms.

Not quite what you would have expected from an old-line chemical company. The point is that technological sophistication is spreading, and as it spreads, it becomes harder and harder for companies that don't join in to maintain a competitive edge. The merging of technologies -- electronics and mechanics, for example, or biotechnology and food processing -- has created the potential for much faster change in industry after industry. Microprocessors are at the heart not just of computers, but also of some of the most advanced locks and ski bindings. General Motors claims to be the largest producer of microprocessors in the world.

Anthropologists say that when one culture invades another, the result, after initial bloodshed and confusion, is often a tremendous flowering of creativity. Something similar seems to happen when technologies collide and coalesce. The challenge for management is to exploit not only the confusion, but also the fusion -- and the energy produced.

Back to top

Sun Microsystems: The New Hotshot

Business Week
February 3, 1986

Summary: The founders of Sun Microsystems Inc. are a study in contrasts -- and in energy that drives a successful computer startup. Andreas Bechtolsheim, a soft-spoken, German-born computer scientist who designed Sun's first workstation as an academic exercise, is prone to kicking off his Birkenstock sandals while pondering a microship question. William N. Joy, Sun's software guru, is an irrepressible extrovert whose bushy hair and beard give him the look of an erudite mountain man. Then there is President Scott McNealy, the buttoned-down yuppie with degrees from Harvard and Stanford, who yearns to play the world's top 100 golf courses.

The founders of Sun Microsystems Inc. are a study in contrasts -- and in energy that drives a successful computer startup. Andreas Bechtolsheim, a soft-spoken, German-born computer scientist who designed Sun's first workstation as an academic exercise, is prone to kicking off his Birkenstock sandals while pondering a microship question. William N. Joy, Sun's software guru, is an irrepressible extrovert whose bushy hair and beard give him the look of an erudite mountain man. Then there is President Scott McNealy, the buttoned-down yuppie with degrees from Harvard and Stanford, who yearns to play the world's top 100 golf courses.

All three are bachelors in their early 30s, married to Silicon Valley's currently hottest startup. They were brought together by a fourth co-founder, Vinod Khosla, an Indian-born entrepreneur who met his goal of retiring by age 30 and is no longer active in management.

The trio built Sun more on technical creativity than discipline, so it has had manufacturing and service problems. But revenues have more than doubled in the last 12 months to $147 million, and McNealy calls $200 million a "reasonable target-- for Sun's fourth fiscal year, ending in June. Hoping to catch rival Apollo Computer Inc. off guard, Sun on Jan. 20 brought out a new, low-cost line that it claims delivers 20% to 33% more processing power than comparable Apollo models -- at half the price.
Apollo executives have tried to ignore Sun's gains and put it down as a price-cutting upstart. But the secret to Sun's success goes beyond the pricing of its machines -- from $7,900 to $50,000.The Mountain View (Calif.) company has moved new technology to market quickly, with an almost fanatical commitment to "open systems" that make its machines adaptable to the industry's evolving technology. Apollo relies on proprietary designs. Sun uses off-the-shelf parts and urges buyers to customize the product.

A lot of customers like that. In early January, Schlumberger Ltd. chose Sun over Apollo for a $65 million contract -- Sun's biggest competitive victory to date. The French electronics and oil-field equipment company wanted "the flexibility of configuring systems any way we want to," says Richard A. Mohrman, president of its Applicon Inc. division. Another big win: a $35 million Toshiba Corp. deal.

Sun's reliance on industry standards is a natural outgrowth of its humble origins. Bechtolsheim was pursuing a doctorate at Stanford University when he set out to design a powerful computer using off-the-shelf components. But he needed money. In early 1982, before he could find a company that would pay to license the design, McNealy and Khosla, classmates at Stanford's business school, suggested starting Sun. They recruited Joy, the architect of an important version of the Unix operating system, which was then gathering momentum as an industry standard for engineering computers. Within a year the four had $4.6 million in venture capital and were shipping $1 million worth of computers a month.
ACHILLES' HEEL. Since then, Sun got an additional $31 million in private financing, including $20 million from Eastman Kodak Co. in a deal that gave Kodak a 7% stake in Sun. Although Sun has been profitable from the start and still has more than $15 million in cash, its growth will demand another infusion of capital soon. "Going public is a distinct possibility," McNealy says.

In preparation for major-league competition, Sun has been hiring better managers. Last year its workstation division recruited Robert Garrow, a co-founder of Convergent Technologies, as general manager. Wayne Rosing, ex-director of engineering for Apple Computer Inc.'s Apple II Group, became engineering manager. In January, Sun raided its archrival for Robert R. Lux, head of Apollo's North American customer-service operations. Lux plans to quadruple Sun's field-service force. This is crucial now that International Business Machines Corp. is entering Sun's market.

These moves should help allay customer concerns over product quality, the Achilles' heel of many startups. The dead-on-arrival rate of Sun equipment "is more than we would like to see," says C. Shelton James, group vice-president of Gould Inc.'s Information Systems Business Section, which signed a $40 million purchase agreement with Sun in May. He adds that Sun "is going through all the things companies go through when they start out."

McNealy insists that quality problems are behind Sun. Now, he's ready to take on IBM. After looking over the giant's latest workstation announcement, he declares with gusto: "We're hungry and looking for lunch."

» Click here for more information

Back to top

Sun's Interest in Stock Offering Is Well-Timed

The Business Journal-San Jose
February 3, 1986
Mountain View, CA

It was just "coincidence" that Sun Microsystems chose to file a preliminary prospectus for an initial public stock offering only two days after IBM Corp. introduced a competing computer. But industry observers said the move was well-timed because IBM's introduction of its RT PC computer workstation cleared up any doubts about its computing power versus similar products from Sun.

"When you look at what the actual (IBM) product is, you can see that Sun has a better workstation," said Sandy Gant, an analyst with InfoCorp, a Cupertino market research firm. "At least in the short term, you look at it and look at Sun, and Sun looks really nice."

Prior to the IBM announcement, Sun would have had to compete with the "myth" of the IBM machine, Gant said. "People didn't know what it was, and they could instill features that it might have. Now, Sun can sell against the reality of the RT."

In a conservative assessment, Sun's prospectus reported, "IBM's entry is likely to make the competitive environment for (Sun's) products substantially more difficult."

Jackie Rae, a spokeswoman for Sun, said it has actually been competing with the IBM RT PC for several months, as IBM showed the machine to some large customers under non-disclosure agreements. Rae said she believed Sun had won most head-to-head competitions with IBM in the race to provide engineers and scientists with computer workstations.

Sun officials were "heartened" by the IBM announcement, Rae said. "We didn't think it was priced as aggressively as it could have been. I don't think we can lose in competition."

Stock analyst Harry Kohli of Sutro & Co. Inc. in San Jose, said Sun "couldn't have picked a better time" to go public. "I like the company. It's very good, technology-wise."

Sun is a leading supplier of workstations based on Unix, the master computer language commonly used in the scientific and engineering communities.

For the six months ended Dec. 27, Sun reported net income of $ 2.7 million, or 11 cents a share, on sales of $ 75.9 million.

Earlier this month, Sun announced a 3-year $ 65 million contract with Schlumberger Ltd. for workstations. Schlumberger reportedly plans to use Sun systems for computer-aided design and engineering, automated testing, scientific research, artificial intelligence and expert systems, and technical publishing.

Sun decided to go public this month for a number of reasons, primarily the stock market's receptivity to high-technology offerings, Rae said. It had begun to prepare an offering last May, but backed off due to a down stock market, she said.

Sources who asked not to be identified said that Apple Computer Inc.'s overtures to acquire Sun or an interest in the company have ended. The San Jose Business Journal earlier reported that executives with the two firms had discussed a merger. The sources, however, said discussions never reached a serious stage.

Sun plans to offer 4 million shares of common stock at an estimated price of $ 16 to $ 18 a share, which would raise $ 64 million to $ 72 million. Among insiders selling parts of their interests, Sun president Scott McNealy plans to sell shares worth as much as much as $ 600,000. As a group, Sun's directors and officers now hold 27.2 percent of the company, which would be reduced to 23.4 percent after the offering.

Sun's largest customer, Eastman Kodak Co., which owns 6.4 percent, would reduce its holdings to 5.7 percent. West Coast Venture Capital of Palo Also, which holds 12 percent, would sell shares worth $ 10 million or more, reducing its ownership to 8.4 percent.

At the same time as Sun registered its offering with the Securities and Exchange Commission, it made an offer to rescind prior sales of $ 3.2 million worth of shares it sold to employees and consultants, by repurchasing them at prices from a third of a cent to $ 2 each, plus interest.

The prospectus says that the sales may have been in violation of federal securities laws.

Sun's Rae said the repurchase offer is being made because the company may have exceeded the number of shares the SEC allows to be sold without filing a registration statement. She said the repurchase offer should satisfy the commission's requirements and allow the offering to proceed without delay due to the earlier sales.

She said it is unlikely any of the shareholders will elect to take the company up on its repurchase offer, since their shares would be tradeable, presumably at a much higher value, on the open market 30 days after the offering becomes effective.

Among facts disclosed in the prospectus, Sun reports that it loaned four of its officers money to buy houses or pay off loans against their residences. They are: a $ 360,000 loan to former Apple engineer Wayne Rosing, who is Sun's vice president of engineering for its workstation division; a $ 380,000 loan to executive vice president Bernard Lacroute; $ 235,000 to James Bean, a vice president; and $ 400,000 to Robert Lux, a vice president.

Sun also reported the following executive compensation for the fiscal year ended June 30, 1985:

Scott McNealy: $ 109,792 in salary; $ 129,956 in bonuses.
Bernard Lacroute: $ 85,000 in salary; $ 37,987 in bonuses.
Robert Smith, chief financial officer: $ 90,000 in salary; $ 38,700 in bonuses.
Joseph Roebuck, vice president of sales: $ 84,000 in salary; $ 33,812 in bonuses.
Darryl Barbe, European general manager: $ 80,000 in salary; $ 24,000 in bonuses.

Back to top

Sun Microsystems growth

Time Magazine
May 26, 1986

Abstract: Sun Microsystems, in a computer industry plagued by layoffs and flat revenues, has achieved phenomenal growth. Sun's specialty is work stations, high-powered desk-top computers that perform as well as much larger and more expensive minicomputers.
ISSN: 0040-781X

Back to top

Rising Sun; Silicon Valley's Hot Newcomer

Time
May 26, 1986
Janice Castro

Summary: A bright glow is rising above Silicon Valley's gloomy horizon. In a computer industry plagued by layoffs and flat revenues, Sun Microsystems of Mountain View, Calif., has achieved phenomenal growth. Founded only four years ago, Sun boosted sales from $8 million in fiscal 1983 to $115 million in 1985. Over the same period, annual profits surged from $654,000 to $8.5 million. Run by Scott McNealy, 31, Andreas Bechtolsheim, 30, and William Joy, 31, a trio of workaholic wunderkinder, the company shows signs of staying power in a business in which success is often fleeting. Sun's specialty is workstations, high-powered desktop computers that perform as well as much larger and more expensive minicomputers. Favored by scientists and engineers, workstations can manipulate graphics and data many times as fast as standard personal computers. In its short life, Sun has captured 20% of workstation sales in the U.S., and is rapidly gaining on Chelmsford, Mass. - based Apollo, which first developed that kind of computer and still holds 39% of the market. The stakes in the competition are enormous: workstation sales are expected to surge from $735 million last year to $2.5 billion by 1989.

A bright glow is rising above Silicon Valley's gloomy horizon. In a computer industry plagued by layoffs and flat revenues, Sun Microsystems of Mountain View, Calif., has achieved phenomenal growth. Founded only four years ago, Sun boosted sales from $8 million in fiscal 1983 to $115 million in 1985. Over the same period, annual profits surged from $654,000 to $8.5 million. Run by Scott McNealy, 31, Andreas Bechtolsheim, 30, and William Joy, 31, a trio of workaholic wunderkinder, the company shows signs of staying power in a business in which success is often fleeting.

Sun's specialty is workstations, high-powered desktop computers that perform as well as much larger and more expensive minicomputers. Favored by scientists and engineers, workstations can manipulate graphics and data many times as fast as standard personal computers. In its short life, Sun has captured 20% of workstation sales in the U.S., and is rapidly gaining on Chelmsford, Mass. - based Apollo, which first developed that kind of computer and still holds 39% of the market. The stakes in the competition are enormous: workstation sales are expected to surge from $735 million last year to $2.5 billion by 1989.

Sun's popular workstations cost anywhere from $7,900 to $70,000, but the price is generally 10% to 20% less than that for comparable models offered by competitors. In addition, Sun users can exchange data with many other types of computers. That allows business customers who already have office computer systems to add Sun workstations. Says McNealy: "Customers can mix and match and create a customized computing environment."

Sun got started in 1982 when Vinod Khosla, a Stanford Business School graduate, brought together McNealy, a former classmate, and Bechtolsheim, a Stanford engineering graduate student who had developed an impressive prototype for a workstation. McNealy recalls that Khosla persuaded him over a dinner of McDonald's Big Macs to quit his job as director of operations at Onyx Systems, a computer company, and help found a new firm. The next recruit was Joy, a Ph.D. candidate in engineering at Berkeley and a leading computer-software designer.

Though Khosla soon withdrew from day-to-day involvement in Sun's operations to realize his goal of retiring at age 30, the remaining trio found that they complemented one another well. Bechtolsheim, who became vice president of technology, was the wizard who designed Sun's early machines.

McNealy, who has been chairman since 1984, is the bottom-line man who hires the employees and makes sure that the products are attractively priced. As vice president of research and development, Joy became the visionary of the group, charged with keeping an eye on the company's future.

Sun made its first public stock offering in March, raising $45 million. When investors eagerly snapped up the stock at $16 a share, the founders, who retain shares worth about $62 million, became overnight millionaires. But they have little time to savor their new wealth. Says McNealy: "We run a nervous operation. We're constantly looking over our shoulder to see who's gaining on us."

One threat is mighty IBM, which has just entered the workstation market with its model RT. To stay ahead of other manufacturers, Sun is boldly aiming to double the speed of its workstations every year.

Back to top

Sun Microsystems introduces latest workstation

United Press International
July 8, 1987
Harihar Krishnan

Sun Microsystems Inc., a fast-rising leader in the computer workstation market, Wednesday introduced its latest supercomputing workstation that promises to significantly improve performance at greatly reduced cost.

Designated the Sun-4 to provide source-code compatibility with the Mountain View, Calif.-based company's earlier Sun-3 and Sun-2 series, the new workstation is a deskside system that can process automated design applications at 10 million instructions per second.

At that rate, Sun claims the Sun-4's speed is more than double the speed of its nearest competitor. The company says Sun-4 will accomplish technical and engineering tasks in minutes that now take hours using mainframes.

The Sun-4's microprocessor or chip is targeted for applications requiring maximum computing power such as computer-aided design, desktop publishing, graphics and image processing, communications and industrial control.

"We expect this product family to redefine workstation computing and create a new price-performance point in the industry," said Bernard Lacroute, Sun's executive vice president.

President and Chairman Scott McNealy, 32, who co-founded Sun Microsystems only five years ago and has seen it grow into a $500 million corporation with 4,000 employees worldwide, said his company "has accomplished something rarely seen in the computing industry by delivering the first supercomputing workstation with a full complement of system and applications software available today."

A Sun-4/260 high resolution monochrome deskside workstation with 8 megabytes of main memory will be priced at $39,900. A Sun-4/260 color deskside workstation with 32 megabytes of main memory, a 560 megabyte disk subsystem and a 60 megabyte -inch cartridge tape system will cost $85,500.
Sample server configurations will range from $36,900 for the Sun-4/260S pedestal model with 8 megabytes of main memory to $104,900 for a Sun 4/280S server with 32 megabytes of main memory and 1.2 gigabytes of disk and tape storage.

The company said the new workstation will give users the performance of a VAX 8800 system at one-tenth the price.

Company officials said the main feature of the Sun-4 series workstation is its scaleable architecture or SPARC, based on RISC (Reduced Instruction Set Computer) technology. This will permit the workstation to be upgraded to significantly increase performance, they said.

"Sun also announced a new server series based on the SPARC technology that offers the highest performance of any UNIX-based system on the market at dramatically lower costs than conventional supercomputers," the company said. "Used as fileservers, compute servers, communication gateways or as cost-effective timesharing systems, these servers are ideal for building highly optimized networks."

The company said it will license SPARC to semiconductor and systems manufacturers and the licensees in turn will supply chips, boards and complete SPARC-based systems to the open market.
Firms already licensed include Fujitsu Microelectronics, Cypress Semiconductor and Bipolar Integrated Technology.

Back to top

Sun's Sizzling Race to the Top

Fortune
August 17, 1987
Stuart Gannes; REPORTER ASSOCIATE Ingrid WickelgrenPicture 1, Money talks: Sun founders Joy, Bechtolsheim, McNealy, and Khosla, with their cars and phones, ED KASHI; Picture 2, To reveal their latest star, McNealy, Joy, and Executive Vice President Bernard Lacroute rented New York's Hayden Planetarium in July. GERD LUDWIG -- VISUM; Picture 3, Strong sales by Carol Bartz's government division mean new plants for Sun in Silicon Valley. ED KASHI

Summary: How do you handle a company whose sales have accelerated from zero to over $500 million in five years flat, especially when you know that a split-second error will mean a fiery crash-- "Like A. J. Foyt," says Scott McNealy, 32, chief executive and a founder of Sun Microsystems, the Mountain View, California, maker of powerful computers called technical workstations. "When two wheels are off the cliff, we start getting careful. With one wheel off, we just step on the gas." Lately Sun has made the competition eat dust. If it doesn't run off the cliff, revenues next year should near $900 million.

HIGHLIGHT:
The five-year-old start-up has roared ahead of the pack in the hot new business of computer workstations. Now it plans to drive into three other markets at once.

HOW DO YOU HANDLE a company whose sales have accelerated from zero to over $500 million in five years flat, especially when you know that a split-second error will mean a fiery crash-- "Like A. J. Foyt," says Scott McNealy, 32, chief executive and a founder of Sun Microsystems, the Mountain View, California, maker of powerful computers called technical workstations. "When two wheels are off the cliff, we start getting careful. With one wheel off, we just step on the gas." Lately Sun has made the competition eat dust. If it doesn't run off the cliff, revenues next year should near $900 million.

Race car imagery is particularly apt in the world of technical workstations. These supercharged machines look like personal computers, but they retail for $5,000 to $100,000 and pack enough power to run dozens of programs simultaneously, display precise graphics, and solve maddeningly complex equations. Originally designed for engineers, workstations are roaring into dozens of government and commercial markets. The industry's sales topped $1.5 billion in 1986 and are rising 30% a year, according to Dataquest, a California-based market researcher.

Sun sells more workstations than anybody else. Its shipments are growing 50% every six months; sales rose 140% to an estimated $517 million in fiscal 1987, ended in June, while profits more than doubled to $37 million. The company's stock, first offered to the public at $16 a share in March 1986, recently traded around $38. Says Mark Stahlman, a security analyst for Sanford C. Bernstein, an investment research firm: "Sun has so much momentum that people on Wall Street are running short of analogies to describe what's going on."

They could soon be speechless. Sun's goal, says McNealy, is nothing less than to become "a broad-based, general-purpose computing company," along the lines of Digital Equipment Corp. He astonished a meeting of security analysts in July by announcing that his company's new Sun-4 workstations could do the jobs of big minicomputers, which accommodate scores of individuals at once, for a fraction of the $700,000 or so that such computers cost. McNealy also hinted that Sun was developing a personal computer to compete with the new generation of machines IBM announced in April. Then Bill Joy, another Sun founder, dropped a third bombshell by predicting that the company would release machines powerful enough to compete with small supercomputers, like those manufactured by Convex of Richardson, Texas, and sold for around $1 million.

Just to stay on its current track, Sun must negotiate some perilous curves. The competition in workstations is getting tougher. Over the next 18 months, Sun will face a field of new, high-end products from powerhouses of technical computing such as Digital Equipment, Hewlett-Packard, and IBM. Upstarts such as Stellar Computer of Newton, Massachusetts, and Dana Computer of Sunnyvale, California, are about to enter the race, as are manufacturers from Japan and Korea. At the low end of the workstation market, which accounts for 25% of Sun's revenues, Apple, Compaq, and others are launching innovative machines that may soon blur the distinction between personal computers and workstations. How can Sun manage to stay on top in its core business and at the same time rush into three new markets without self-destructing?

Sun has performed wonders in the past. The company's first product was the brainchild of Austrian-born Andreas Bechtolsheim, a shy graduate student in electrical engineering at Stanford University. A poet of circuit diagrams -- friends say he's more comfortable talking with transistors than with people -- Bechtolsheim built his workstation out of spare parts scrounged from Silicon Valley supply houses. In 1982 he teamed up with three other 27-year-olds, all newcomers to the computer business. Two were Detroiters: the ebullient, can-do McNealy, son of an American Motors vice chairman, and software genius Joy, then a frustrated grad student at the University of California at Berkeley. India-born Vinod Khosla, McNealy's roommate at Stanford Business School, wrote Sun's first business plan (see box). The name Sun is an acronym for Stanford University Network, the communications project for which Bechtolsheim designed his workstation.

The four neophytes faced a powerful competitor. Apollo Computer of Chelmsford, Massachusetts, pioneered the business, launching its first workstation in 1980, a year and a half before Sun came to market. Apollo quickly dominated the infant industry, and it has performed sensationally, racking up revenues of $392 million last year and profits of $9.3 million. The seven-year-old company is a sure bet to make the FORTUNE 500 next year. But so is Sun, which shot past Apollo in the March quarter with $18.3 million more in revenues than its older competitor.

The secret of Sun's growth: defying conventional wisdom. The only way to crack the engineering market, said the experts, was to build a unique product whose inner workings were based on proprietary designs. Although the research and development expenses for such a machine would be enormous, owning the technology would ensure fat profits. The master of the strategy was DEC, whose scientific computers came to populate nearly every research lab in the free world. Apollo also executed a unique-product plan with textbook precision.

Sun's founders believed that a workstation built from off-the-shelf components -- everything from chips to basic operating software -- would attract plenty of customers and be cheaper to produce. Instead of matching Apollo's prices and going for high margins, McNealy and his crew decided to cut prices and concentrate on building market share.

THE STRENGTH of Sun's approach, says Bill Joy, is that it recognizes a central truth: Technological change in the computer industry is continually accelerating. No single company can be at the forefront of every important breakthrough, and those that try doing it all themselves -- the classic vertical integration route -- inevitably fall behind. Cooperating with the technology leaders, Joy maintains, is the only way Sun can grab a permanent place out front.

The strategy has a glaring weakness. By depending on off-the-shelf hardware and software, Sun leaves itself wide open to competitors, from a more efficient manufacturer to one with more marketing clout. Sun's defense is to keep the competition perpetually off balance with a barrage of new products. Manufacturers of mainframe computers typically upgrade their products every four to five years, and personal computer models tend to stay on the market for two years or more. But Sun issues forth workstations every 12 months on average, nearly as fast as Sony comes up with new versions of the Walkman. Each new Sun offers at least twice the power of its predecessor for roughly the same price. Says Vice President Carol Bartz, who runs Sun's government business: "We wouldn't hesitate to bring out a new product at a price and performance level that absolutely destroyed an existing line. Why should we wait for the competition to do it-- That's a brand-new concept in this business, and we've proved you can make money doing it."

Sun's adherence to industry standards of hardware and software delighted computer buyers, who fear locking themselves into a proprietary technology. To make its systems even more alluring, Sun devised a communications network that allows customers to attach Sun workstations to any other computer. McNealy, a supersalesman, came up with the line: "In the past, computer companies have been able to charge a premium for proprietary technology. In the future they will have to offer a discount."

The message soon struck pay dirt. Big customers such as the engineering departments of Boeing and General Electric, which ordinarily disdain buying from unproven start-ups, ordered truckloads of Suns. Today the average Sun salesman brings in nearly $3 million a year in revenue, more than twice the industry average -- all without significant advertising. Says Bartz: "The reason we got past $500 million in five years was that we were a safe buy. And that was because our products were based on standards."

Selling standards is still fueling Sun's growth, especially in markets where workstations are just catching on. One of the biggest is Washington. Says Bartz: "The federal government is the largest user of computers in the world, and its appetite for workstations is just overwhelming." Sun's government sales, primarily to the military and the intelligence community, tripled to $30 million last year. One federal spy agency recently ordered Sun products worth more than $200 million.

WALL STREET is as willing as Washington. Sun's workstations are sprinkled across the trading floors of Morgan Stanley, L.F. Rothschild, and several other firms. Financial consultants such as the Athena Group of New York are using Suns to develop artificial intelligence programs that will forecast market movements by analyzing thousands of transactions. In fact, Sun's machines have become a Wall Street status symbol. Earlier this year Kidder Peabody recruited two "rocket scientists" -- technical analysts -- by promising them their own machines. "They told me they wouldn't have come unless we bought them each a Sun," says Paul Sobota, a Kidder vice president.

Sun's victories are transforming the workstation business. Today all its competitors offer machines that run the Unix operating system, an off-the-shelf set of programs for controlling a computer's basic functions that Sun has used from the start. "Every computer company eventually will have to make a major commitment to Unix, from supercomputer makers like Cray on down," says Apollo founder John William Poduska, who now heads Stellar Computer, a start-up workstation company. Apollo is also saying "me too." At its most recent product announcement, CEO Tom Vanderslice declared: "We will offer standards wherever they exist."

Even Digital Equipment, potentially Sun's most powerful opponent, is changing its tune. "Sun has been successful in selling to many of our customers. We need to respond with something better," says Christopher Reed, head of workstation marketing for DEC. In June, barely four months after introducing its latest workstations, DEC slashed their prices more than 50%, making some of them even cheaper than Sun's.

Despite his upbeat demeanor, McNealy spends a lot of energy imagining the worst, saying: "We've studied every crash and burn in the industry, and we don't want to repeat any mistakes." He runs Sun like a commander on red alert, constantly requesting detailed updates on production, sales, and hiring. Monthly financial reports run to 200 pages, "and you'd better be sure you know what's on page 53," says Vice President Robert Smith. Gauging the firepower of the competition is the responsibility of marketing director John Hime. "I keep a two-year forward set of charts, in excruciating detail, of what every one of our major competitors is going to offer," Hime says.

THOUGH SUN is confident it can handle everything the competition dishes out for the next year and a half, it is not content to coast along. Its new Sun-4 workstations use microprocessor chips of Sun's own design. Unlike standard products from Motorola and Intel, these chips are high-performance thoroughbreds. Designed according to the principles of RISC -- reduced instruction set computing -- Sun's micro is far simpler than standard chips because it processes fewer basic instructions. Simplified RISC chips can be made by almost any kind of semiconductor manufacturing method without being redesigned. As chipmaking techniques improve, Joy says, Sun will be able to double its microprocessor's performance every year through 1990. The chip is key to Sun's ambitious plans to become a full-line computer company.

With its RISC chip, Sun is trying to create a new standard. Instead of buying the best available components, it will get the best available chipmaking technology. Sun cannot afford to make chips, so it is licensing the RISC design, along with necessary software, to several outsiders. Sun will get the microprocessors it needs for a broad range of machines at very low cost, and the chipmakers will acquire valuable new products to sell. Japan's Fujitsu, the seventh-largest chip producer in the world, was the first to take up Joy's offer in 1985. A year later two other leaders in chip technology, Cypress Semiconductor and Bipolar Integrated Technology, also signed on.

To be sure, Sun's RISC strategy entails special dangers. If the licensees fail to sell the chips to lots of other computer makers, the micros will not become standards and Sun's customers may start to worry about being locked into a proprietary design. By licensing programs it has written to run on the chip, Sun is effectively selling off one of its most valuable assets. Any competitor, from a hungry start-up to an established giant, can do to Sun precisely what Sun did to the rest of the industry over the past five years. In June and July, as Wall Street analysts tried to assess the impact of the company's RISC strategy, Sun's stock bounced wildly between $35 a share and $46.

Will Sun burn itself out by offering RISC as a standard-- Joy thinks not: "Just watch a great chef prepare a meal from scratch. There's a whole lot more to cooking than buying ingredients, and a lot more to computers than buying hardware and software." McNealy is less certain. Says he: "We always said, 'Let's think big.' If we fail, we'll make the biggest crash anyone has ever seen. But if we succeed, we'll change the fundamentals of the computer business the same way Henry Ford changed the fundamentals of the automobile business."

Back to top

Sun Microsystems may set standard with chip design; Sun Microsystems Inc. is believed to be negotiating to license its proprietary SPARC reduced instruction set computer chip design to several firms

PC Week
October 6, 1987
Russell Glitman

Sun Microsystems May Set Standard With Chip Design

Wall Street analysts believe Sun Microsystems is negotiating to license a proprietary RISC processor-chip design to a number of firms, including AT&T and Intergraph Inc.

"There are a number of companies evaluating the chip design," said Peter Rogers, an analyst with the New York investment-banking firm Mabon Nugent. "Which and how many will bring products to market, however, won't be known for another year."

The Sun chip, called SPARC, a reduced-instruction-set chip (RISC), allows for faster processor performance. Fujitsu is producing a 10-million-instruction-per-second (MIPS) version under a Sun license. A 16 MIPS version should be available from Cypress Semiconductor in San Jose, Calif., at year's end, and a 50 MIPS version is under development at BiPolar Integrated Technology in Beaverton, Ore.

The Fujitsu chips go into Sun's newest workstation, the Sun/4, as well as into machines made for Prime Computer, Mr. Rogers said.

AT&T's Place

According to a New York-based analyst, who asked not to be named, licensing the chip design could give AT&T its best chance to enter the commercial workstation market and to stake out the high ground in microcomputing.

"There's a lot of logic to something like that happening," the analyst said. "AT&T is clearly moving toward high-performance machines," he added, citing the recent introduction of the firm's mid-range 3B400 Apache. At the introduction of that product, AT&T President Vittorio Cassoni said the machine is the perfect product to support RISC processors.

"AT&T is still trying to find a place in the computer business, and the workstation area is the highest-growth segment of the market," Mr. Rogers noted.

Analysts agreed that license fees from the chip will have little direct impact on Sun's bottom line. But should a wide range of workstation vendors embrace the chip, it would catapult Sun into a standards-setting position, they said.

"It's a major plus for Sun each time another systems vendor signs up for the Sun SPARC," said the New York analyst.

AT&T and Sun have a long-standing cooperative relationship. The two are currently working on the second version of Sun's operating system--a converged operating system that incorporates AT&T's Unix System 5 and Berkley 4.2.

Aggressively marketing the chip would fall in line with Sun's corporate strategy, which to date has been based on implementing standards, Mr. Rogers said.

To maintain that strategy with a proprietary processor design, Sun and its chip makers will "have to be able to license it to the rest of the world," he explained.

But one West Coast analyst doubts AT&T and Sun are in talks over the use of the chip in future products. Richard Edwards, an analyst with the San Francisco investment banking firm Robertson Coleman Stephens, said, "I don't give it a lot of credence."

Back to top

Bill Joy: Outspoken UNIX Guru

Lotus
April 1988

Summary: This past January, Apple Computer and Digitial Equipment Corp. announced an alliance that some analysts believe will put additional pressure on IBM. Sun Microsystems' cofounder and vice president of research and development, Bill Joy, responded, "So what--"

"I assume DEC will say 'this is wonderful' and Apple will say 'together we are better than IBM'," says Joy. "My take is that it isn't good enough. Apple doesn't know how to build large computers, and DEC isn't good at the desktop." It certainly isn't in Joy's interest for Apple and DEC to leave the rest of the industry in a cloud of dust. But Joy's views are important, considering what he has accomplished at Sun.

This past January, Apple Computer and Digitial Equipment Corp. announced an alliance that some analysts believe will put additional pressure on IBM. Sun Microsystems' cofounder and vice president of research and development, Bill Joy, responded, "So what--"

"I assume DEC will say 'this is wonderful' and Apple will say 'together we are better than IBM'," says Joy. "My take is that it isn't good enough. Apple doesn't know how to build large computers, and DEC isn't good at the desktop."

It certainly isn't in Joy's interest for Apple and DEC to leave the rest of the industry in a cloud of dust. But Joy's views are important, considering what he has accomplished at Sun.

Sun Microsystems (Mountain View, Calif.) has been growing at a record pace. Last year Sun had sales of $ 537 million, making it the leading manufacturer of the powerful desktop computers used primarily by scientists and engineers. Since its founding in 1982, Sun has made networking and open systems based on industry standards essential ingredients of its recipe for success.

"Suns work fine with IBM machines, Macs, and others because Sun machines were designed to work with other machines," Joy says, adding, "We're committed to communicating with everybody else, and I don't have to make a big announcement."

Joy's company has made its share of noise recently, however. Most of it has resulted from an agreement under which AT&T will acquire as much as 20% of Sun. Industry experts say the accord could catapult Sun into a position of direct competition with IBM and DEC in the market for general-purpose office computers. The foundation for Sun and AT&T is the Unix operating system and a new microprocessor design developed by Sun.

A brilliant programmer, Joy is recognized as one of the computer industry's great Unix gurus. He's also arcitulate, and his sense of humor shows as he describes single-vendor, proprietary standards (such as those offered by Apple and DEC), which he feels are against the best interests of the customer, as "on the dark side of the Force."

While a graduate student at the University of California, Berkeley, Joy adapted AT&T's Unix, and the result has become a standard among researchers nationwide. It also helped prolong the life of Unix and extend its uses beyond the laboratory. Now Sun and AT&T are working on a new Unix version for the next generation of desktop computers.

The false starts Unix and AT&T have encountered so far in the business world don't discourage the ever-optimistic Joy. "Unix is a great boon to creativity," he says. "It solves a lot of problems. There's no reason you can't run applications software on Unix or any other operating system."

As for AT&T's repeated failures to enter the computer business, Joy shrugs, "Bell Labs invented the transistor and the laser. They built a hell of a phone system. They were constrained by the government from entering computing and took a while to adjust. They're a great partner. They get more than three strikes. They get as many at bats as they want. I think they'll do well this time."

Some industry experts see the AT&T/Sun mix as a gamble. For Bill Joy it fits neatly into his view of computer technology, which is known by his colleagues as Joy's Law.

According to Joy's Law, the speed of computers will continue to double each year. By the end of the century, we will use systems that are a fraction of the size of today's systems but have 128,000 times the power. That belief drives Joy.

"I guess I'm an eternal optimist," he says. "I think we can double the power of the machine each year. You're on the curve or you're off the curve." To reach his goal, Joy believes the computer industry must focus on open standards instead of engaging in parallel research and development efforts that result in worthless, private standards. "It's important to have competitors building compatible products," says Joy. "It keeps people honest and it makes more value for the customer," he adds.

But not of all of Joy's competitors share his philosophy. In fact, some view the specter of Sun and AT&T as a direct threat not only to IBM but to other hardware manufacturers as well. As of this January, more than a dozen of them, including DEC, Hewlett-Packard, Apollo Computer, and NCR Corp., have met to discuss the impact of a new version of Unix to be developed by Sun and AT&T. These companies fear that Sun and AT&T will optimize the operating system for Sun's new SPARC chip at the expense of other hardware makers.

"If we're in a position where Unix is controlled by one of the competitors in the business, that is completely unacceptable," says Larry Lytle, a manager of consultant and analyst relations at Hewlett-Packard.

The critics charge Sun with playing a game of calling for open standards while at the same time developing technology that will work best on its own technology.

Joy was surprised by the industry reaction and insists his goal is simply to improve Unix, noting that he wants to make public all forthcoming changes to the operating system. "We're going to give the code to everyone. We want Unix to succeed," he says.

Back to top

Fun Days at the Sun Frat House

Fortune
May 23, 1988

Summary: Last April Fool's Day, Scott McNealy, the 33-year-old founder and chief executive of Sun Microsystems, drove to work and found his office transformed into a miniature golf course. The night before, a band of engineer pranksters had torn out the wall behind McNealy's desk, removed the furniture, and covered the floor with fresh sod. The one-hole course they built was a tricky par 4. From an elevated tee, the fairway took a wicked dogleg to the right. Extra hazards included two sand traps and a birdbath, not to mention the glass picture window. Total distance: 12 yards.

Last April Fool's Day, Scott McNealy, the 33-year-old founder and chief executive of Sun Microsystems, drove to work and found his office transformed into a miniature golf course. The night before, a band of engineer pranksters had torn out the wall behind McNealy's desk, removed the furniture, and covered the floor with fresh sod. The one-hole course they built was a tricky par 4. From an elevated tee, the fairway took a wicked dogleg to the right. Extra hazards included two sand traps and a birdbath, not to mention the glass picture window. Total distance: 12 yards.

McNealy was so flabbergasted that he bogeyed the hole. "I went into the trap, and it took me two shots to reach the green," he explains.

Such outrageous stunts are not only tolerated by Sun's exuberant leader, they are savored. In fact, with his buckteeth and boyish looks, McNealy seems more like the head of a college fraternity than the chief of a FORTUNE 500 company. And freewheeling Sun, nestled in the heart of Silicon Valley, does nothing to dispel the image. In addition to its tradition of yearly April Fool's shenanigans, the company stages weekly dress-down days and throws monthly beer bashes. On Halloween, employees show up for work in gorilla suits. Says McNealy: "We're trying hard to be different from other companies. One of our goals is to provide an environment that people have a blast working in. We like to think of Sun as a billion-dollar startup."
Fair enough. At the age of six, Sun can hardly be called mature. Nor can McNealy, who came to Sun just two years out of business school. But the young chairman's personality is a perfect match for Sun's wide-open culture. Under McNealy's leadership, Sun raced ahead of its competitors to become the leading manufacturer of high-performance technical workstations -- the current rage among engineers, scientists, and financial analysts, who crave the extra horsepower these desktop computers offer. For Sun's fiscal 1987, which ended last June, revenues increased 156% to $538 million. And sales in its most recent quarter set a $1-billion-per-year pace.

Sun's growth astonishes McNealy as much as anyone else. Raised in Detroit, he is the son of an American Motors vice chairman. McNealy studied economics at Harvard and earned a business degree at Stanford in 1980. One of only three members of his class who sought manufacturing jobs, he joined FMC in Chicago. Within ten months he quit to work for Onyx Systems, a Silicon Valley computer maker. "FMC put me on a strategy team, and I wanted to be a plant manager. I wanted to make something," he recalls. McNealy left Onyx in early 1982 when a business school buddy, the brilliant India-born Vinod Khosla, asked him to help start Sun. Khosla was Sun's first chief executive. McNealy was head of manufacturing. "That meant I built the first 25 Suns by hand," he recalls, "and I had the skinned knuckles to show for it."

McNealy was named Sun's interim president in 1984; the man the company's board of directors had recruited for the job, a veteran executive from Digital Equipment Corp., never fit in with Sun's rambunctious troops, and resigned. A few months later the 30-year-old Khosla decided to retire, and McNealy was appointed interim chairman as well. "At some point, I guess I became more than interim," he says.

WITH 6,500 EMPLOYEES, Sun is no longer a mere startup, and McNealy is growing with his job. Over the past year he has proved himself a tough and forceful executive as he hammered out major deals with such industry giants as AT&T, Xerox, and Unisys. Yet McNealy still relishes his role as Sun's most enthusiastic cheerleader. "I'm not a strategizer at heart," he says. "I'm more focused on cohesion and pulling everybody together. Goals only limit you.We let the market and our ability to have fun set the company's goals."

McNealy's business heroes come from the computer and car industries. They include Kenneth Olsen, founder and chief executive of Digital Equipment, as well as William Hewlett and David Packard, the pair of billionaires who inspired thousands of Silicon Valley entrepreneurs by starting their company with $538 in a garage. "These are guys who stuck with their strategies for a long, long time, all the while protecting the cultures of their corporations," says McNealy. He reserves a special affection for Henry Ford. "By bringing automobiles down the cost curve, Ford really changed the way we live our lives," he says. "The same thing is going to happen with computers."

Still a bachelor, McNealy plays hockey at an indoor rink in San Jose and golf on any course he can find. "My real dream is to play enough golf to make it to the Masters," he jokes. Give his engineers the word and they just might decide to stage that hoax, come next April.

Back to top

They Are Not There Waiting for the Plane to Land

Has Sun Microsystems, one of the computer industry's most spectacular successes, now taken on more than it can handle?
Forbes
June 27, 1988
Kathleen K. Wiegner

Summary: Just don't use the words 'brash' and 'arrogant' when you talk about us," jokes Scott McNealy, president of Mountain View, Calif.-based Sun Microsystems, Inc. Unfortunately for McNealy, Sun's reputation in the computer industry -- for an aggressiveness bordering on hubris -- is no laughing matter. Six-year-old Sun's jet-propelled rise (revenues should be close to $ 1 billion this fiscal year, ending June 30) has won it high praise on Wall Street and in the press. But even in a highly competitive industry, Sun is unusually friendless. Its one good buddy, AT&T, is not exactly helping: Its relationship with Sun has drawn the fire of some of the biggest guns in the computer industry.

HIGHLIGHT:
Has Sun Microsystems, one of the computer industry's most spectacular successes, now taken on more than it can handle?

JUST DON'T USE the words 'brash' and 'arrogant' when you talk about us," jokes Scott McNealy, president of Mountain View, Calif.-based Sun Microsystems, Inc. Unfortunately for McNealy, Sun's reputation in the computer industry -- for an aggressiveness bordering on hubris -- is no laughing matter. Six-year-old Sun's jet-propelled rise (revenues should be close to $ 1 billion this fiscal year, ending June 30) has won it high praise on Wall Street and in the press. But even in a highly competitive industry, Sun is unusually friendless. Its one good buddy, AT&T, is not exactly helping: Its relationship with Sun has drawn the fire of some of the biggest guns in the computer industry.

Almost everyone has a snide "Sun story" about the company's high-handed manner and prideful claims that only Sun gives customers what they really want. Take the matter of how well Sun's computers perform. One joke competitors like to tell compares Sun's performance claims to whipped butter -- 50% air.

Until now, having built a billion-dollar company in just six years, McNealy could have been entitled to shrug off such sniping. But he has attracted a crowd of enemies so large and so powerful that the climate at Sun is about to become a scorcher.

The company picked up its first $ 1 billion as one of the pioneers of desktop workstations -- a combination of hardware and software offering the power of a minicomputer but with the easy desktop access of a personal computer. Comparing a workstation to a personal computer, at least until recently, was like comparing a Ferrari to a Chevette. Chip designers doing custom designs, manufacturers doing solid modeling, chemical companies doing molecular analysis and even the quants on Wall Street found these workstations just the ticket. Estimates are that this year technical workstations, including those from such companies as Apollo, Digital Equipment and Hewlett-Packard, will constitute a $ 2.8 billion market.

But the distinction between a workstation and a personal computer has become less clear since Sun introduced its first product in 1982. Until recently personal computers generally processed data 16 bits at a time while workstations did it in 32-bit chunks. And personal computers used an Intel microprocessor and the MS-DOS operating system developed by IBM and Microsoft while Sun workstations used an operation system called Unix, developed by AT&T, which was better than MS-DOS for networked computers. Unix not only handled certain complex operations better but it was a so-called open system in that it did not compel its users to buy only one brand of equipment.

Then, in 1986, Intel brought out its 32-bit microprocessor, the 80386, which has given PC makers enough computing power to blast into Sun territory. Compaq Computer, for example, estimates it currently sells 40% of its high-end, 386-based personal computers into the technical marketplace, a Sun stronghold. Many of these computers use Unix rather than the MS-DOS operating system that is standard on most personal computers. Estimates are that between 15% and 20% of Apple Mac II's are also being sold with Apple's version of Unix rather than with Apple's proprietary operating system.
This spread of Unix was not lost on its developer, AT&T. Estimates are that by 1991 perhaps 20% of the computers sold will use Unix. So last fall AT&T turned to Sun -- whose co-founder, William Joy, is considered one of the best Unix engineers in the business -- for help in developing its next version of Unix. Then in January of this year AT&T announced that it planned to buy up to 20% of Sun over a three-year period (it currently has a 7.1% stake in the company).

All this understandably alarmed the other companies that sell Unix-based computers, particularly when they heard that Sun's salesmen were suggesting to customers that Sun would not have an inside track on future Unix developments. Hewlett-Packard, one of the best-mannered companies in the industry, was rebuffed when it tried to be included in the AT&T/Sun Unix development team working in Menlo Park, Calif.

So in May seven computer makers (who in aggregate represent some 40% of the computer industry), including not just H-P but IBM and DEC, announced they were banding together in a nonprofit organization called the Open Software Foundation to develop their own "open" version of Unix that would not be "proprietary" to AT&T and Sun. If the industry comes to view the AT&T/Sun version of Unix as proprietary, it may become very interested in the more "open" system being proposed by the Open Software Foundation.

If the Open Software Foundation comes up with a product in a timely fashion, IBM would be the big beneficiary, since its version of Unix, called Aix, will be a core technology in the Foundation's development plans. But even if it doesn't, the move creates confusion in the market that is likely to delay software development. The company with the biggest sales force, the saying goes, has the best chance of "unconfusing" a customer. That distinction does not belong to Sun.

Sun has compounded its problems by trying to push the industry into adopting as a standard the innovative RISC (reduced instruction set computer) microprocessor called Sparc, which it had designed. The problem-- The computer companies are intensely suspicious because, unlike a Motorola or an Advanced Micro Devices, which also makes RISC chips but well only chips, Sun sells computers. If Sparc became a standard, competitors fear, Sun would have the inside track on new products containing the chip. The result-- Do not bet heavily that the industry will move to a Sparc standard.

Finally, Sun has now decided to attack the personal computer guys head-on, particularly Apple and Compaq. In April Sun introduced its 386-MS-DOS as well as Unix, to compete directly with the top end of Apple's and Compaq's lines. This battle could get bloody. Apple and Compaq are already well established in the retail channel. Sun, in contrast, has relied on its crack direct sales force, which is traditionally the way technical computers were sold. "We all agree that workstations and personal computers are on a collision course," says Enzo Torresi, vice chairman of Businessland, the computer retailer (see story, p. 58). "The question is, who will get there faster-- Distribution is the key, and Sun doesn't have the distribution that Apple and Compaq do."

Any one of these fights would be tough enough for a seasoned computer company. But Sun's youthful management (average age, just over 42) is now stretched thin as Sun tries to execute all its strategies at once. Some technical types have begun whispering that Sun has lost its technical edge and is simply trading on its reputation and market size. "Sun seems to be able to handle its manufacturing, procurement and shipments," says one disinterested observer, "but its R&D and marketing are in chaos. They have a lot of momentum but are not well disciplined."

For Sun's current fiscal year, revenues could reach $ 1 billion, almost double 1987's revenues. Earnings are estimated at $ 1.57 to $ 1.60, up from $ 1.11 a year earlier. But its aggressive pricing strategies and its willingness to pay premium prices to secure scarce 1-megabit D-RAMS (FORBES, June 13) have lowered its gross margins. Should revenue growth slow substantially, margins would be too slim to prevent a big earnings hit.

McNealy thinks his company is now big enough to weather a mistake or two. "At $ 1 billion in revenues," he says, "we can stub our toe and not crash and burn." But he is also well aware that should the world not turn out to be heliocentric, no one will weep for his misfortunes. "There is definitely a crew waiting at the end of the runway," says McNealy, "and they are not there waiting for the plane to lane."

Back to top

Sun Microsystems Turns on the Afterburners

Business Week
July 18, 1988
Jonathan B. Levine

Summary: It was a big day at Sun Microsystems Inc. On a June afternoon, 2,000 employees from the burgeoning Workstation Div. gathered beneath a circus big top to celebrate their graduation to group status. Onstage, group executives sported black gowns and mortarboards. At the head of the line was Scott G. McNealy, Sun's 33-year-old president, grinning behind a pair of his trademark Ray Ban sunglasses.

HIGHLIGHT: It's churning out products at a faster and faster pace. Too fast?

It was a big day at Sun Microsystems Inc. On a June afternoon, 2,000 employees from the burgeoning Workstation Div. gathered beneath a circus big top to celebrate their graduation to group status. Onstage, group executives sported black gowns and mortarboards. At the head of the line was Scott G. McNealy, Sun's 33-year-old president, grinning behind a pair of his trademark Ray Ban sunglasses.
Comedian Pat Paulsen, the hired entertainment, quickly caught the mood. "What's the difference between Sun and the Boy Scouts--" he deadpanned. "The Boy Scouts have adult supervision." The audience roared as Paulsen redoubled: "Who's running this place anyway, Beaver Cleaver--"

The party just gave observers more evidence of the youth, exuberance, and brass that have made Sun Microsystems a billion-dollar company in six years, one of the hottest computer stocks on Wall Street, and one of the most controversial new competitors in its industry. But suddenly its success, so rapid and sweeping, is sowing seeds of doubt among an increasing number of industry watchers. No one's predicting a precipitous fall. But no one's completely at ease either. Increasingly, the question is asked: Is Sun spinning out of control--

While the company is pursuing everything from its own microchip design to minicomputer-class machines, competitors are closing in on its main business, high-powered workstations for engineers and programmers. An alliance with American Telephone & Telegraph Co. has provoked Sun's four biggest competitors -- IBM, Digital Equipment, Hewlett-Packard, and Apollo -- into joining forces against it. The result, says Goldman, Sachs & Co. analyst John C. Levinson, is that "a lot of people are waiting for Sun to blow up." Even McNealy concedes: "I don't know of any other company our size facing the sort of challenges we face."

Few computer startups have come as far as Sun without major setbacks. Major crises, such as Apple Computer Inc.'s loss of founder Steven P. Jobs, seem almost a rite of passage in Silicon Valley. McNealy and his managers have avoided them so far by eschewing the hierarchical structures and bureaucracy that slow down other organizations. Now, says Robert M. Kavner, president of AT&T's Data Systems Group and a Sun director, "it's getting to the size where it needs business discipline -- and yet not the bureaucracy."

Sun's challenge, then, is to instill order without dousing the passion. But Sun executives have little time to ponder technique. The company has prospered by building computers that embody widely accepted standards such asEthernet networks and AT&T's Unix operating system, the software that controls the basic functions of a computer. Because such standards made it relatively simple to switch to another make, customers were willing to gamble on the young company. Standards were also a big advantage over rival Apollo Computer Inc., whose proprietary operating system made it difficult to change brands. Now, however, all Sun's rivals sell Unix workstations, too. To compete, "we have to be faster out the door and better than everyone else," says Bernard J. Lacroute, Sun's executive vice-president.

'CONTROLLED CHAOS.' Sun is certainly trying. Its 7,000 employees are encouraged to share ideas, usually via the company's electronic mail system. Major decisions, sometimes arbitrary but always quick, reflect a consensus of senior executives who meet in noisy, table-pounding meetings. Product strategy bubbles up from autonomous divisions rather than from a central committee. As Vice-President Eric E. Schmidt puts it: "Sun is controlled chaos."

Is this any way to run a $ 1 billion company-- Given Sun's view of the industry, it may be the only choice. McNealy believes a shakeout among computer makers is coming by the early 1990s, and that Sun's best chance to survive is to grab as much market share as possible now. That explains the rapid proliferation of new products. In addition to workstations -- powerful desktop computers geared to technical tasks -- Sun now makes systems that can be shared by a department full of workers the way DEC's VAX minicomputers can. There also is a low-end workstation similar to the most powerful IBM and Compaq personal computers. Rumored projects range from workstations for university students to mainframe-size systems.

The pace has veteran industry watchers panting. "Sun's not only trying to grow faster than any other company, but it's trying to do it with one of the broadest lines," says Laura Conigliaro, an analyst at Prudential-Bache Securities Inc. "It's unprecedented." It's also expensive: Sun spends about 14% of revenues on research and development vs. the industry average of 8.2%. That and aggressive pricing have helped keep net margins a modest 6%.

For now, rapid growth takes precedence. Although he recently lumped nine divisions into two groups, McNealy prefers to keep business units small and agile. Three years ago he began breaking Sun into autonomous divisions to focus on specific markets, such as high-performance workstations and systems sold to government agencies. He even funded an independent East Coast Div. for personal computers. So far the groups have thrived by doing things their way. In its third year, ended June 30, the federal systems operation had sales of $ 120 million. The East Coast Div., which shipped its first product in April, is exceeding the $ 250 million annual sales rate the company forecast.

The key is getting individuals to take on enormous responsibility with little direction from managers. For instance, in 1986, Apollo introduced software that threatened Sun's dominance in workstations used by programmers. That business accounted for 40% of Sun's sales, so it kicked into overdrive, turning 10 engineers loose on the project. In nine months -- half the normal development cycle -- they had an improved programmer package that blunted Apollo's threat. "Nothing motivates Sun like the fear of what a competitor might do," says Wayne E. Rosing, vice-president for advanced development.
It's a good thing. In the past year, Hewlett-Packard Co. and Apollo have regained momentum in the $ 3.6 billion workstation market, and DEC has publicly predicted that it will unseat Sun as No. 1 this year. Meanwhile, Silicon Graphics Inc. and startups are challenging Sun technologically in areas such as 3-D graphics. Powerful personal computers such as Apple's Macintosh II and IBM's Personal System/2 are encroaching from below. And although Sun has comparably priced machines, these competitors have an advantage that Sun is just now trying to match: thousands of computer dealers to blanket the market.

With competition like that, Sun more than ever needs 110% output from its employees. One motivating force is stock. Through a stock-purchase plan and bonuses, 80% of employees own Sun shares, which have more than doubled since the company's 1986 initial offering. Even so, revving up newcomers is a chore. Nearly half of Sun's employees were hired within the past year, and the company is still adding 300 more each month. To speed acculturation, each new hire is assigned a "Sunvisor," an experienced colleague who knows the ropes. This spring the company started its first formal management training program. "Our biggest challenge is getting mid-managers as gung-ho as those of us who've been here for years," says Carol A. Bartz, head of federal systems.

POINT MAN. Except for Vinod Khosla, who dropped out of daily operations amid political frictions early on and resigned as a director last year, the founders are still hard at work. William N. Joy and Andreas V. Bechtolsheim remain the company's technical gurus. But Chief Executive McNealy, a Harvard graduate and Stanford MBA, is the point man. The son of an American Motors Corp. vice-chairman, he grew up in a family whose social connections include the likes of Lee Iacocca and Roger Smith. So when it came to dealing with venture capitalists and Wall Street bankers or negotiating with AT&T, "I never felt intimidated," he says.

Although a persistent salesman with outsiders, within the company McNealy goes to great lengths to build a consensus among his executives. "Give me a draw and I'll make the decision," he says. "But I wouldn't edict something on an 11-to-1 vote." By now, Sun's management philosophy has become an amalgam of traits brought by top managers from DEC, Apple, and Intel. Lacroute, 45, is an especially significant import. After 14 years at DEC, where he managed a division, he's the guy who keeps track of daily operations. Nicknamed "Little Napoleon" for his diminutive stature and hard-nosed style, Lacroute is the levelheaded alter ego to McNealy that one-time Apple President Mike Markkula was to young Steve Jobs.

Even with Lacroute's influence, however, the sheer chutzpah of McNealy and his crew has outraged most of the computer industry Establishment. Sun raised hackles last summer when it proffered its own Sparc microchip as an industry standard to be used by competitors in their computers. Relations with other computer makers became more strained last fall when AT&T agreed to use Sun's microchip in its minicomputers and gave Sun the job of co-developing an improved version of Unix. Other suppliers of Unix systems thought that might give Sun and AT&T the edge in bringing out computers that could use the latest version of Unix. In particular, they worried that Unix would be altered to run most efficiently on Sparc.

CRYING FOUL. Relations with the competition deteriorated after AT&T announced plans to acquire as much as 20% of Sun and named a director to its board. Then the two companies proposed that Sun's Open Look "graphical user interface," a program to give Unix workstations the type of screen format popularized by Apple's Macintosh, be used as an industry standard. Although AT&T plans to license Open Look as part of Unix next year, competitors cried foul. Apollo Chairman Thomas A. Vanderslice claimed that Sun and AT&T were giving themselves as much as an 18-month advantage in bringing Open Look products to market ahead of other suppliers of Unix computers. Within a month, Vanderslice and the heads of DEC, IBM, HP, and several other computer makers had banded together to announce the Open Software Foundation (OSF), a consortium dedicated to producing an alternative to the AT&T-Sun version of Unix.

Although OSF's own operating system is more than a year away, the rift could be troublesome for Sun. Confusion over which operating system will prevail could slow development of applications software for Sun's machines and hamper efforts to have its Sparc chip incorporated in more computer brands. That would blunt Sun's efforts to expand beyond engineering and into commercial data processing.
To some observers, the Unix controversy seems the inevitable consequence of being too cocky. "A lot of Sun's newer managers think they're infallible," says co-founder Khosla. "And that's the first step toward fallibility." The company's growth hasn't slowed, but there are trouble spots ahead as Sun attempts to compete more directly with IBM and DEC. The scramble to get hot products to market quickly has already produced three different varieties of Sun hardware, making it more difficult to switch software from one Sun machine to another.

Another weak spot is service and support. Sun's haste to grow has left it little time to develop the type of bedside manner that IBM and DEC offer large customers. This weakness may be particularly harmful as Sun tries to win commercial customers, and some technical types already consider it a problem. Dissatisfied with Sun's aftersale support, Schlumberger Ltd.'s computer-aided design division is replacing half its 200 Sun workstations with DEC machines. "They don't have the expertise in the field," says Terry L. Dollhoff, a Schlumberger vice-president. "There are some chinks in Sun's armor."

Investors aren't worried. Although Sun continues to sacrifice margins for growth, its stock, currently close to 40, trades at 20 times calendar 1988 earnings -- higher than most of its peers' shares. McNealy also has learned to anticipate the concerns of jittery investors. He was the first major computer executive to warn analysts to cut forecasts because of the memory-chip shortage. So when Sun outdid expectations, most were pleasantly surprised. "At 33, I can't afford to appear out of control," says McNealy. And at 6, neither can his company.

» Click here for more information

Back to top

Sun's President Talks About the Chip Shortage, Degree of OSF's Impact

PC Week
November 14, 1988

Summary: Sun Microsystems Inc. President Scott McNealy is making faces at his vice president of marketing, trying to get him to crack up. The boyish 34-year-old McNealy can afford to joke around. His firm is on a roll. Sun recently ended its sixth year by breaking the billion-dollar sales barrier. Now McNealy, the son of a Detroit auto executive, is a Silicon Valley celebrity. McNealy spoke with PC Week at Sun's headquarters in Mountain View, Calif. After the interview, McNealy was off to a kids' grand prix speedway, which Sun had rented to honor five-year employees. "Only one guy beat me at the last one," he said.

Back to top

Sun aims revamped OS at Intel arena; SunSoft Inc. to offer SunOS under the name Solaris; multiprocessing version planned

PC Week
September 2, 1991
Laura Brennan and Andrew Ould

SunSoft Inc., the Sun Microsystems Inc. subsidiary formed to sell system software, will make a splash this week with two new versions of the SunOS operating system, now renamed Solaris.

Hoping to make Solaris a star among corporate PC users, SunSoft officials are expected to announce at Sun's developer's conference this week in San Jose, Calif., that they will port a version of the enhanced software to Intel Corp.'s 32-bit processors, said sources close to Sun. The software was primarily designed for Sun's SPARC platform.

Solaris 1.0, a replacement for the existing SunOS 4.1, will include Sun's OpenWindows windowing scheme, a new Desk Set suite of office-productivity tools and Sun's Open Network Computing protocols, the sources said. The package is due in the fourth quarter, they added.

The second version, Solaris 2.0, will merge 1.0's features with AT&T's Unix System V release 4.0 and will support multiprocessing for Sun's forthcoming multiprocessor server, code-named Galaxy, sources said. It was unclear last week when Solaris 2.0 would be shipped.

Both Solaris versions will offer binary compatibility with SunOS 4.1, enabling users to run Sun applications unmodified. DOS emulation will also be available for DOS users, according to sources.
Analysts said porting Solaris to Intel platforms will create a new business opportunity for Sun, a $ 3 billion workstation maker. "Sun is differentiating itself from competitors through software, where the profit margins are high," said Larry Frietag, an analyst at Standard & Poor's Corp., an investment-rating firm in New York.

Sun spokesman Zach Nelson, in Mountain View, Calif., declined to comment on specifics but said it's "certainly within SunSoft's charter to look at other architectures and port software to them." In April 1988, Sun tested the Intel market with its 386i workstation, but discontinued that machine late last year to focus on SPARC.

Sun's latest move to Intel platforms also presents an attractive alternative to the Advanced Computing Environment (ACE), analysts said. Companies backing ACE plan to standardize on Intel-based PCs and workstations based on MIPS Computer Systems Inc.'s RISC chips. Both platforms will run Microsoft Corp.'s Windows NT operating system and The Santa Cruz Operation Inc.'s version of Unix. There is one snag, however: Microsoft's Windows NT is not available yet.

"If Solaris runs on the 486 and 586, do you need NT on a MIPS chip-- Maybe the whole theory behind the ACE alliance is moot," said Richard Shaffer, publisher of Technologic Partners' Computer Letter, in New York. "Sun has the potential to get the jump on NT."

Apart from the Intel port, analysts said Solaris is also notable because of its compliance with AT&T's standard Unix.

"Right off the bat, at least 18,000 or 19,000 applications written for V.4 will run unmodified on Sun's new operating system," said Rikki Kirzner, an analyst at market researcher Dataquest Inc., in San Jose, Calif.

Back to top

Sun to Enter Race to Offer Software

The New York Times
September 4, 1991
Andrew Pollack
San Francisco

Sun Microsystems Inc. is beginning an attempt to become a major supplier of operating system software for computers, putting it into direct competition with Microsoft and a new joint venture of Apple and I.B.M.
Sun will announce on Wednesday that it is developing a version of its operating system, which is the basic set of commands for controlling computers, that will run on personal computers using microprocessors made by the Intel Corporation. Sun's operating system now runs only on machines that use Sun's Sparc microprocessor. Such machines are mainly Sun's own work stations.
The announcement by Sun is the latest in a series of scrambles by companies for position in the computer industry of the 1990's. Many companies are realizing that the key to profits and control in the industry is not so much making computers, but rather providing software that can control many different types of computers.

Moves by Others

The Microsoft Corporation, whose MS-DOS operating system controls most personal computers, is developing new software, known as Windows NT, that will run both on Intel-based personal computers and on machines using a microprocessor developed by MIPS Computer Systems. Apple Computer and the International Business Machines Corporation have announced their intention to co-develop an advanced operating system that will run on several different types of microprocessors.
"Sun and Apple in the last several months looked in the mirror and said, 'We are a software company,' " said Andrew Grove, president and chief executive of Intel.
With its announcement today, Sun, the fast-growing leader in the engineering work-station business, is hoping that its operating system will establish its position before the other new systems are developed.
"Here we have a product that we think is head and shoulders above what they are talking about," said Edward J. Zander, president of Sunsoft, a software subsidiary formed by Sun in July.

Based on Unix

The Sun operating system, known as Solaris, is based on the American Telephone and Telegraph Company's Unix operating system. Sun said the version of Solaris for Intel microprocessors will be available by the middle of 1992.

By making Solaris run on Intel microprocessors, Sun hopes to broaden the market for the operating system and attract software companies to develop programs for it. Some software companies have been hesitant to develop programs, like word processors and spreadsheets, for Sun work stations because Sun sells only 250,000 machines a year, compared with millions of Intel-based personal computers sold each year. But now, software companies will be able to develop one version of their programs that will run both on Sun machines and on many personal computers.

A common operating system would allow Sun machines to blend better into offices that have personal computers, helping Sun expand its market beyond engineers. The risk for Sun, however, is that making its software available for Intel-based computers could reduce the attraction of machines based on its own Sparc chip.

Indeed, the announcement will represent a triumph for Intel, whose dominance of the microprocessor business has been threatened by chips like Sparc. But Intel is drawing encouragement from the fact that Solaris, as well as the new operating systems being developed by Microsoft and by I.B.M. and Apple, will all run on its chips.

Continuing in the recent flurry of coalition-forming in the industry, Sun, which is based in Mountain View, Calif., has enlisted help from other companies.

Novell Inc. of Provo, Utah, a leading provider of networking software for personal computers and a rival of Microsoft's, will distribute the Solaris product and help make it work together with its own software. The Dell Computer Corporation of Austin, Tex., a major manufacturer of I.B.M.-compatible personal computers, will offer Solaris with some of its machines, and AST Research Inc. of Irvine, Calif., another major clone manufacturer, will certify that Solaris will run on its computers.

Back to top

Will Sun Also Rise in the Data Center?

June 1, 1993
Datamation - Scott McNealy

Back to top

Why Sun thinks hot Java will give you a lift new software designed to make world wide web's 'home pages' more useful; and spur computer sales

San Jose Mercury News
March 23, 1995
David Bank

A few months ago, establishing a "home page" -- or site -- on the Internet's World Wide Web was enough to win a company or individual a place among the information cognoscenti.

Now, many of multimedia's hippest designers deride today's Web pages as static, boring and dumb.
A home page is a computer file with text, photos or graphics that can be viewed via the World Wide Web and can contain links to other files.

Many leading-edge designers today are buzzing about Sun Microsystems Inc.'s new software that the Mountain View-based company hopes will turn the Web into a rocking new medium. The software enables producers to make the Web as lively as a CD-ROM, but with the added advantages of continuous updates and real-time interaction between people.

"Rather than just having a CD-ROM that gets pressed once and is immediately out of date, you can interact with people on the Net," said Karl Jacob, the chief technologist at Dimension X Inc., a San Francisco game producer that is using the software to construct "environments" for multiple computer users to visit, such as virtual saloons, dance halls and poetry- reading salons.

"What we're trying to do with this is build a community around the site," Jacob said. "You don't just download pictures and text that just sits there. You can interact with characters on the screen or play games or browse 3-D environments."

Next week, Sun plans to release an early version of the software, known as Hot Java, to give multimedia developers a chance to create attractions in time for a wider release, which is scheduled for June.

In the tradition of the Internet, Sun plans to give the software away for free. The first release of Hot Java will work only on Sun's workstations, which means nearly no one will be able to use it from their home computers. By June, the company said, versions will be available for computers running Microsoft's Windows 95 and the new operating system for Apple Computer's Macintosh.

Even competitors are impressed.

'Undeniably, absolutely new'

"What these guys are doing is undeniably, absolutely new. It's great stuff," said Marc Andreessen, vice president for technology of Netscape Communications, which has by far the largest share of the market for current- generation World Wide Web browsers. "There's so much stuff that people want to do over the network that they haven't had the software to do. These guys are really pushing the envelope."

In a demonstration this week at Sun's lab in Palo Alto, Hot Java's product manager, Kim Polese, showed off a financial planning application with a ticker of selected stocks scrolling across the top of the screen with up-to-the- minute quotes. On the same page, the changing stock prices were rendered in graphic format.

On another page, Duke, a red-nosed, molar-shaped imp that is Hot Java's mascot, performed cartwheels and backflips in a looped animation.

Behind the presentation was a new software programming language developed by Sun that changes the relationship between the computers that send the information -- known as servers -- and those that receive the information - clients.

How it works

Hot Java takes advantage of the computing power of the client computers. Instead of simply downloading static pages of text and photos, the Hot Java browser downloads small software programs, which then run on the client.

That makes Hot Java fast and versatile, but more importantly, extremely adaptable.

"The place you can enhance the users' experience the most is right in front of them, but the browser right now is the dumbest part," said Patrick Naughton, who helped develop Java for Sun before joining Starwave Corp. in Seattle as vice president for technology, where he is using Java for on-line services with ESPN and Outside magazine.

He said that with Hot Java, fantasy sports players will be able to see complete, up-to-date statistics for every player and game, make trades in real-time and follow their favorite teams with customized scorecards and news. Other applications

In Outside Online, a mountain biker who selected a particular trail map could chat on-line with others in the same area, trading information about trail conditions, directions and hazards. Such live chats are a staple of on- line services and other parts of the Internet, but have not been possible over the World Wide Web.

"It's a kick start for a new category of applications," said Eric Schmidt, Sun's chief technology officer. "A whole lot of companies, including companies we don't even know about, will invent brand new ways of doing interactivity on the Net."

Java is the result of six years of work, at least that many name changes and millions and millions of dollars. Development started in 1989 with a super-secret project called Green, which was seeking a way to control consumer appliances such as toasters and light switches.

Analysts who have previewed the software said that even now Sun may not be able to create a viable business around Hot Java.

Sun: It's an investment

Schmidt said that Sun was not worried about making money on Hot Java directly, but would profit if the software aided the overall growth in the use of the Internet. Schmidt said the ability to create innovative services on the Internet would spur information and service providers to buy more Sun workstations, which are among the most widely used servers on the Web.

"This is an expense that we bear to make the network more successful," Schmidt said. "Sun is in the business of having the inter-networking pie get bigger."

Back to top

Sun Microsystems Creates New Program Language for Web

The Associated Press
May 23, 1995
New York

Sun Microsystems Inc. on Tuesday introduced software that will give the World Wide Web portion of the Internet flexibility to exchange data that is more complex.

Sun's Java programming language and HotJava browser for moving around the Web are designed to do more than just read documents known as "Web pages."

Web pages typically use text and graphics, which are fairly easy to manipulate. That's important because the Web's popularity is due to the simplicity to jump between Web pages that are stored on thousands of different computers.

The Sun products make it possible to download small software programs, allowing the Web documents themselves to have more features. Programmers, for example, could build games that are played simultaneously by many people or financial programs that are updated with stock or mutual fund prices.

HotJava also contains a number of security features that verify information and protect against viruses and tampering, Sun said.

The company, a maker of workstations and server computers, will distribute the Java software and related HotJava browser free for non-commercial use. While early test versions are available now, a more complete one is not expected until later this year.

Back to top

Sun's Rise

BusinessWeek
January 22, 1996
Robert D. Hof

SCOTT McNEALY'S RISING SUN
How he's taking the computer maker to new heights

Scott McNealy was dog-tired. For weeks, the chief executive of Sun Microsystems Inc. had been sealing deal after deal for Java, Sun's red-hot Internet software. By the first week in December, a bunch of the top names in computing--everyone from Internet star Netscape Communications to database kingpin Oracle Systems to IBM--had endorsed Java. Everyone, that is, except Microsoft Corp. The software giant was holding out for the same reason everyone else was so excited: Java programs run on any hardware or operating system, bypassing Bill Gates's cash cow, Windows.

After three frenzied days of meetings in New York, McNealy was looking forward to a full night's sleep. All the schmoozing--including talks with IBM CEO Louis V. Gerstner Jr. that had yielded a promise by the world's biggest computer maker to use Java--was taking its toll. And back home, McNealy's 4-week-old firstborn had been cutting into his shut-eye. The following day, Dec. 7, would be another marathon: an interview at CNBC, customer meetings, and a trip home on the leased corporate jet in time for a business dinner--and a San Jose Sharks hockey game.

NO LIGHTWEIGHT. At 2 a.m., the phone in his room at the Grand Hyatt jangled McNealy awake. It was Sun co-founder and Javameister William N. Joy with big news: Joy had just received a fax from Microsoft Senior Vice-President Roger Heinen. Microsoft had agreed to license Java on Sun's terms--ending four weeks of negotiations. Time to uncork the bubbly-- Or spread the word across the Net-- No way. Chronically sleep-deprived, McNealy mumbled, ``Great job," and rolled over.

Correction: incredible job. Microsoft's capitulation stands as a defining moment for Sun and its 41-year-old CEO. For more than a decade, McNealy has shouted himself hoarse preaching the gospel of network computing--the notion that you begin to realize the true value of computers only when they work together in networks. The message worked well with technically demanding customers in engineering departments and Wall Street trading rooms, who couldn't survive on ordinary personal computers and their crude local-area networks. So Sun became tops in engineering workstations. But it barely made a dent in the far larger market for mainstream office computers. By the early 1990s, sales slowed, profits dropped, and it seemed the company would never break into the big time.

That, however, was before the Internet explosion and Java. With businesses tripping over themselves in the past 18 months to get on the Net, Sun's mantra--``The Network is the Computer"--has begun to resonate around the globe. Even the mighty William H. Gates III--whose software dominates the huge PC market--concedes that McNealy's Sun has the right stuff for the next era of computing.

That Sun now finds itself in the center of the computing universe has a lot to do with the unlikely character at the top. The toothy, boyish-looking McNealy has an unusual resume in an industry driven by technical whizzes and creative entrepreneurs. Unlike Gates or Steven P. Jobs, he didn't drop out of college to answer the call of the burgeoning PC business. Nor did he work his way up through engineering, as did Intel Corp.'s Andrew S. Grove and dozens of other Silicon Valley execs.

You won't find McNealy sweating over computer code down in the engineering lab. He studied economics at Harvard University and got his MBA at Stanford. His passions are hockey and golf. And he's famous for sophomoric pranks and shoot-from-the-lip one-liners. Typical of his wit is this assessment of Microsoft's Windows and MS-DOS: ``whipped cream on a road apple."

But McNealy has proved to be anything but a lightweight. Indeed, while CEOs with better technical credentials have flamed out--flitting from one techno-fad to another or neglecting the fundamentals of the business--McNealy has emerged as one of the Valley's most respected managers. Programmer-turned-entrepreneur Lawrence J. Ellison, CEO of Oracle, salutes McNealy's blend of talents. ``There are two things I think about Scott," says Ellison. ``One is passionate leadership, and the other is his rigorous financial management. And that's uncommon to find in one person. Usually, the financial guys aren't so outspokenly passionate, and all leaders are not detail-oriented."

Those talents, plus a killer competitive instinct and nonstop drive, have kept Sun on course through a decade of wrenching change in the computer industry--while IBM stumbled and new leaders such as Intel and Microsoft emerged. Now, Sun has a shot at the leadership in the network era.

TOY STORY GLORY. As McNealy has long predicted, the network has won. The typical stand-alone PC, he insists, has been a complex time-waster, a ``hairball on the desktop." Now, thanks largely to the Net, consumers and businesses have come around to Sun's view: When computers are networked, their power multiplies geometrically. Not only can people share all that information inside their machines but they can reach out and instantly tap the power of other machines--essentially making the entire network their computer.

Sun didn't exactly invent network computing, but it can claim to be the only pure play in the business. The two technical forces behind Sun--co-founders Bill Joy at the University of California at Berkeley and Andreas V. Bechtolsheim at Stanford--had been computer-science students at a time when many of the networking technologies that formed the basis of the Net were created. They built that technology into Sun's first computers in 1982 and into every Sun machine since.

Now, all this is paying off. Already, approximately 35% of all the World Wide Web servers in the world are Sun machines. And as companies adopt Internet standards for internal networks, they are turning to Sun. Federal Express, Gap, AT&T Universal Card Services, and Charles Schwab all use Sun networks. And thanks to the new UltraSPARC chip, Sun's workstations are once again on the leading edge. Sun machines recently stole some of Silicon Graphics Inc.'s Hollywood pizzazz by helping to create the blockbuster movie Toy Story.

In Sun's first fiscal quarter ended Sept. 30, earnings jumped 120%, to $85 million. That followed an 82% hike in fiscal 1995, to $356 million, on a 26% jump in revenue, to $6 billion. Sun's stock tripled, from 16 to 45 5/8 since last year, adjusted for a 2-for-1 split on Dec. 12. But like other technology companies, Sun's stock has been giving back recent gains, plunging 6.8% on Jan. 9, to 37 3/8. Typically, McNealy isn't satisfied, complaining that other companies--including newcomer Netscape--rate higher price-earnings ratios. His New Year's resolution: a $100 stock price by yearend.

Pretty heady stuff. ``We're all having a blast," McNealy says. But nobody's goofing off, least of all McNealy. He's painfully aware that this golden moment will soon pass. ``I just wish I didn't have to sleep," he says.

In the meantime, Sun still faces big challenges. For one, despite the buzz about Java, the company still has to figure out how to make it a business (page 73). And there's a downside to the triumph of Sun's network vision: Every computer maker now shares it and is gunning for Sun. At the same time, Intel and Microsoft continue to push PC performance into workstation territory, and Sun's workstation rivals are coming on strong. What's more, the company's ascent into the top tier of computer makers puts new demands on Sun. In addition to pumping out killer hardware and software, it also must cater to the needs of big customers--the way rivals IBM, Hewlett-Packard, and Digital Equipment long have done.

That's why, after a short Christmas break with his family--his wife of 16 months, Susan, and son, Maverick (the name will be redundant, boasts Dad)--McNealy is back to his 80-hour weeks and heavy travel schedule. ``We have too much work to do," he says.

Besides, he loves the pace. In fact, he was born to it. His father, R. William McNealy, was vice-chairman of American Motors Corp. By the time he was a teenager, Scott was spending evenings with his dad, poring over AMC memos and golfing in foursomes with such industry luminaries as Lee A. Iacocca. Scott says he saw how AMC was marginalized because it never had sufficient market share--and he has vowed to make sure that doesn't happen to his company.

At Cranbrook Kingswood School, a prep school north of Detroit, McNealy emerged as a leader, captaining the tennis team. At Harvard, he became interested in economics--thanks in part to class section leader William J. Raduchel, who would later become McNealy's chief information officer and corporate development chief at Sun.

But McNealy was not a standout student. He was rejected from both Harvard and Stanford business schools--three times from Stanford. So he took a job as foreman at a Rockwell International Corp. plant in Ashtabula, Ohio, in 1976. The company was rapidly building truck hoods in expectation of a strike. The grind was so hard--two months of 14-hour days--that McNealy says it led to a bout of hepatitis that kept him hospitalized for six weeks.

``KIND OF A GOOF-OFF." When he finally did get into Stanford, his study habits and choice of curriculum didn't scream ``CEO material." For one thing, he focused on manufacturing at a time when finance was the route to the top. Moreover, classmates say McNealy preferred playing golf to attending classes. ``He was kind of a goof-off," says Curt Wozniak, a Stanford buddy and former Sun engineering vice-president. McNealy says he just didn't bother attending classes that he didn't think would help him get a job. ``I minimized hours per grade point," he boasts. After graduating, he took undistinguished manufacturing jobs at tankmaker FMC Corp. and minicomputer maker Onyx Systems.

In those days, McNealy seemed determined not to pursue his dad's CEO quest. He once vowed never to work as hard as his father: That workaholism eventually led McNealy's parents to divorce. Instead, McNealy dreamed of running a small machine shop he could give to his kids and retiring early.

Then, in 1982, came the call that would change McNealy's fate. Former Stanford classmate Vinod Khosla asked McNealy to join him and computer designer Bechtolsheim in starting Sun. McNealy's manufacturing skills enabled the young company to keep up with wild demand as sales soared from $9 million in 1983 to $39 million in 1984. McNealy says he got manufacturing humming so well that soon he wasn't getting enough orders to sop up production. So he started running sales. But the new orders outstripped the cash available for expansion. So McNealy was tapped to look for help--and he settled on customer Eastman Kodak Co., which was using Sun workstations in a microfilm project.

McNealy's tenaciousness and quick thinking impressed J. Philip Samper, then executive vice-president at Kodak. As a condition of investing $20 million, Kodak insisted McNealy take over as president in 1984. Samper later served as president of Sun's hardware unit. At about the same time, then-CEO Khosla left in a dispute with the board, so McNealy was appointed president, at least until the board found someone with more experience. But Sun's fortunes rocketed, and the directors stopped looking for a replacement. At 30, McNealy was officially named CEO.

McNealy soon etched his image on the industry. Although he says he dislikes being described as brash--his tag in dozens of newspaper and magazine stories--brash he was. In 1987, Sun joined with AT&T to develop a new version of the Unix operating system. IBM, DEC, HP, and other computer makers struck back by forming the Open Software Foundation to create another version of Unix. Never mind that these were the biggest companies in the industry and Sun was a relative upstart. McNealy publicly mocked OSF and its members, roaring that the initials really stood for ``Oppose Sun Forever."

At Sun's Mountain View (Calif.) headquarters, he was building a corporate culture based on his own motto: ``Kick butt and have fun." The company has become equally famous for its aggressive marketing and the juvenile antics staged around its headquarters. Each April Fools' Day, scores of photographers now arrive to record the elaborate pranks Sun engineers play on McNealy and other execs. Once, the company's engineers built a golf course hole in McNealy's office, complete with green and water hazard.

McNealy likes to join in the high jinks, too. He has played general at an intramural squirt-gun war. Last May, at a conference in San Francisco, McNealy spent a half-hour earnestly laying out Sun's vision for 4,000 computer buyers and programmers. Then he set out several cardboard fire hydrants bearing the names of rivals and introduced Network, a Greater Swiss Mountain dog that stars in Sun's ad campaign. Sure enough, McNealy led the dog to the hydrants and exhorted him to raise a leg. Network, evidently more tactful than his master, declined.

The humor, which is turned on everybody, including McNealy himself, has an important effect: It binds the company and helps employees live with their demandingjobs. ``His humor and ability to raise a crowd to its feet is in many respects exactly what you need in CEOs and leaders of today's industry," says Thomas J. Meredith, a former Sun treasurer and now vice-president for finance at Dell Computer Corp. McNealy's special gift, says former sales vice-president Carol A. Bartz, now CEO of Autodesk Inc., is to energize his people. ``Energy comes right out of his pores," she says.

NEWT BOOSTER. These days, McNealy tries to channel some of that energy to his new family. In addition to Susan and Maverick, the family now includes Network, McNealy's own Greater Swiss Mountain puppy. Although his compensation hit $3 million last year and the 2.9% of Sun shares he owns is worth over $200 million, McNealy remains firmly planted in the three-bedroom house he bought in the hills north of the Valley in 1986, although he owns a condo in Palm Springs. He still likes nothing better than playing hockey and is on two league teams so he can always find a game. Whether it's on the ice or the links, McNealy is determined to win. ``It's just not in my nature not to be competitive," he concedes.

For socializing, he likes to barbecue and sit around with friends, drinking domestic beer and watching a hockey game on TV. Says Dick Boyce, a Stanford mate and PepsiCo Inc.'s chief financial officer: ``He would be happy to be described as Joe Six-Pack working hard to get the job done."

His politics, however, are anything but mainstream. He's a libertarian, who has been known to rail against all forms of government regulation and the welfare state, and he's a big fan of Newt Gingrich. That's unusual in Silicon Valley, where top executives turned out for Bill Clinton, but it doesn't stop the affable McNealy from mixing with other Valley notables. Lately, in fact, he has been spending time with Steve Jobs--Laurene Jobs and Susan McNealy met at an engagement party for McNealy given by Ellison. Now, the two families--both have infants--get together occasionally for dinner.

Almost nothing, however, takes McNealy away from Sun for long. ``There was full disclosure" about that before the wedding," he says. ``I'll never know my kids as well as most people," he adds wistfully. ``But I've also got 15,000 people and $7 billion worth of market value counting on me."

Sun is not a one-man show. ``One thing Scott has done is he has supplemented his leadership with some wonderful people," says Ellison. ``You don't find Scott surrounded by dummies. You find Scott surrounded by real smart people, like Bill Joy and Eric Schmidt [chief technology officer] and others who do wonderful work." Those others include Edward J. Zander, president of Sun's hardware unit and the key marketing executive, and Raduchel, who as chief information officer is responsible for getting the company to run on its own machines.

To make the growing organization more nimble, in 1991, McNealy split Sun into seven ``planets"--for hardware, software, service, etc. The setup got off to a rocky start because of turf battles among executives. Finally, in the summer of 1993, the CEO issued an ultimatum, says Schmidt: ``Sign up, or get the hell out." Now, McNealy says his managers still aren't getting along as well as they should, but at least they're no longer fighting over who sells what.

Concerned about his effectiveness, McNealy recently hired a ``CEO coach," Chuck Raben of Delta Consulting Group Inc. Raben polled Sun managers on where the boss could improve. The consensus: McNealy, who admits he has an opinion on everything, needs to listen better. Raben gave McNealy a cheat sheet, reminding him, for instance, to make sure he has covered all the points raised in meetings with managers.

Still, Sun's execs don't want McNealy to change too much. ``Management is always contentious at Sun," says Raduchel. ``Conflict-free management is always a disaster."

A POST-WINTEL ERA-- And nobody wants McNealy to abandon his strong opinions. After all, the Java technology that has made Sun the darling of the Web exists today in large part because of McNealy's stubborn belief in it. In late 1990, McNealy asked a departing engineer, Patrick Naughton, to take a last moment and write a memo on how Sun could improve. Naughton suggested creating a software system for portable devices--and McNealy persuaded him to stay and do it.

McNealy gave the project, code-named Green, total independence and served as its cheerleader. Once, with the team burning out, he dropped in for a demo. It was a shaky prototype, but the boss turned morale around when he exclaimed: ``This is the greatest thing I've ever seen," Naughton recalls.

Eventually, Green dead-ended with projects for interactive-TV set-top boxes and personal digital assistants. But even as Green wound down, says Naughton, now vice-president for technology at multimedia startup Starwave Corp., ``Scott salvaged a lot of things" --one of them Java. Says pal Steve Jobs: ``When he believes in a strategy, he will stick with it through thick and thin."

An example of that is McNealy's determination to continue to do it all--chips, software, hardware, the works. That has been increasingly risky as off-the-shelf commodity technology has swept the market. But McNealy is convinced that the Net--and Java, in particular--will alter the dynamics of the business. ``Applets" written in the Java language can be zapped across the Net to run on any device that has a Java ``virtual machine"--a tiny bit of code that mimics a computer. If Java catches on big, the software lock-in of the Microsoft Windows/Intel (Wintel) design will end, McNealy says. Then, computer and software companies will once again be able to differentiate their products. Indeed, they'll have to. ``If you're going to make automobiles, you have to make your own engine," he says.

Before that post-Wintel era arrives, McNealy has plenty of other battles. HP, IBM, DEC, and Silicon Graphics are all gunning for Sun's core computer business. HP claims to have stolen workstation business from 35 Sun customers since 1993 by offering what Sun concedes is superior service and support of big customers. ``They're losing in workstations, playing catch-up in servers, and have lost their lead in the overall Unix market," says HP Senior Vice-President Willem P. Roelandts. Some customers worry that Sun has too much on its plate to focus on service deficiencies. ``They're grabbing at everything," says Stephen W. Butt, a manager at a Lockheed Martin Corp. unit that switched to HP workstations.

Sun has beefed up its $885 million service and support unit, and upped the head count 50%, to 3,000 since 1994. McNealy, who calls on customers whenever he can, keeps close tabs. At Schwab, for instance, he checks in every six weeks or so. ``Scott has followed up religiously," says CIO Dawn Lepore.

It's worth the effort. Commercial computing accounts such as Dunlop Tire Corp., which used Sun servers to replace mainframes, now make up a third of Sun's revenues. And the servers they buy help the bottom line: While growth in the workstation market has slowed and margins have contracted, the higher profits from servers have helped Sun boost its corporate gross profit margin to 44% for the first quarter, ended Sept. 30, from 41% in fiscal 1994.

By the end of 1996, Sun may find it more difficult to get such profits from hardware. That's because of Intel's next-generation Pentium Pro chip. PCs that use it will rival workstations in power. Worse, Intel has a four-chip ``motherboard" that will let companies such as Compaq Computer Corp. crank out servers costing a quarter as much as Sun's midrange machines. ``Intel servers will be the dominant servers on the Internet," predicts PaineWebber Inc. analyst Stephen K. Smith.

McNealy's answer, again, is Java. Once the new software succeeds in undoing the Wintel lock, customers will see how outstanding Sun hardware and software are, he argues. ``Java opens up a whole new world for Sun," he says.

This view of the futureis not a universal one. Despite all the excitement, Java is unlikely to transform computing overnight. First, there are security issues raised by a system of distributing software on the Net. And for now, the Java programs run slowly. ``Java is still radically immature," says independent software consultant Mark Pesce.

This bothers McNealy not a bit. ``I want Sun to be controversial," he says. ``If everybody believes in your strategy, you have zero chance of profit." McNealy does fret that with so much on his plate, he'll be blinded to the next big thing: ``We're so frantic with everything going on right now that we're missing something." This season, though, there's little danger of a solar eclipse.

Back to top

Sun introducing new line of servers, targeting mainframe market

Associated Press
January 22, 1997
Mountain View, CA

Sun Microsystems Inc., faced with growing competition for its lower-priced business machines, introduced a line of higher-priced servers today.

The new line of servers - powerful machines that manage computer networks - is intended to compete against mainframe systems.

The Ultra Enterprise 10000 servers will start shipping in limited quantities this week and in volume in March. They will start at $ 873,890 apiece.

Sun is the leading maker of workstations, powerful desktop computers used by researchers and engineers. But in recent years it has faced competition from personal computers.

But among Sun's other products are servers, which are increasingly important for businesses. Sun now hopes to win the business of corporations using mainframes, which can take up entire rooms and cost millions of dollars and are difficult to upgrade.

Hewlett-Packard Co., Digital Equipment Corp., and IBM already make high-end servers.
Shares of Sun Microsystems slipped 12 cents to $ 33.25 in morning trading on the Nasdaq Stock Market.

Back to top

JAVAMAN THE ADVENTURES OF SCOTT MCNEALY TODAY'S EPISODE HIS FIGHT TO SAVE THE WORLD WIDE WEB FROM THE EVIL EMPIRE

Fortune Magazine
October 13, 1997
Brent Schlender

It does sound like the plot of a comic book, except that there are billions of dollars at stake, not to mention maybe the future of business and consumer computing. Standing between all of us and utter subjugation to Bill Gates and Microsoft, there remains just one stubborn obstacle: an $8.6-billion-a-year maker of powerful server computers and software called Sun Microsystems. It's a face-off that pits Scott McNealy, Sun's pugnacious, shoot-from-the-lip CEO--who at times seems more like a trash-talking two-dimensional comic book figure than a real captain of industry--against the richest and still-getting-even-richer man in the world.

Bear with us for a moment; there's a business point here. McNealy is armed with computerdom's buzziest technology--an Internet-friendly computer language called Java--and he's backed by a rabble of computer and software heavyweights including IBM, Oracle, Netscape, and Apple, all of which share to some degree his fear and loathing of Microsoft. Their goal for Java: to render Microsoft's flagship Windows operating systems insignificant--or at least less almighty--with what they believe to be a software lingua franca. The much hyped Java digital language would enable all computerized devices, not just orthodox Wintel PCs, to share programs and commune over a network.

McNealy is best known for virulent, colorful, often highly ad hominem criticisms of what he calls "Bill Gates' centrally planned economy." At a recent public appearance, he snidely referred to the Microsoft founder and his chief lieutenant, senior vice president Steve Ballmer, as "Ballmer and Butt-head." For better or worse, the snottiness is not likely to change. "I cringe sometimes with the way he says things," sighs Bill Raduchel, Sun's chief information officer and McNealy's mentor since college. "But now he's trapped. If he makes a moderate statement, people expect more."

Still, McNealy has earned prominence through real accomplishments, not just hyping Java and bashing Bill Gates. Under his leadership, Sun has cashed in better perhaps than any other company on the business world's networking appetite. Sun's potent lineup of high-end computers for networks--"Unix servers" in nerdspeak--today delivers power that Intel PC servers outfitted with Microsoft's Windows NT are only beginning to approach. Indeed, McNealy, who has been CEO for all but two of Sun's 15 years, has built an efficient powerhouse that has more than held its own against larger, entrenched adversaries such as IBM, Hewlett-Packard, and Digital Equipment. Despite doomsayers' predictions, Sun also has so far nimbly sidestepped the Wintel PC juggernaut.

To learn where Sun stands in the industry, read the following story, "Meanwhile, Back at Headquarters." This story is about Javaman, the manager. There's more to him than his mouth. He's complex, smart, and fiercely ambitious--a 3-D character who is gunning to become as influential an industry leader as his nemeses, the Lords of Wintel.

THE FACTORY RAT

I love the factory. That's what business is all about. Making things. Not this Wall Street stuff, not this consulting stuff, not this lawyer stuff. None of that adds value. You gotta make something. --Scott McNealy

They don't call computers "boxes" for nothing. People are far more interested in the geeks who design them than in how they are actually made. Yet if there is a key to McNealy's success as an executive, it is his Rustbelt upbringing and his factory man's passion for getting product out the door. Says Eric Schmidt, Sun's former chief propeller-head who is now CEO of Novell: "Scott is not an intellectual. He's a driver who realizes that sometimes it's more important to execute well than to have the strategy right the first time. He gets that from his manufacturing background. I find myself trying to copy many of his methods."

McNealy got a feel for nuts and bolts at a very early age. His father was an up-and-coming manager for American Motors who made his mark revamping the parts-procurement systems at various plants. As a grade-schooler, Scott would sometimes abscond with his dad's briefcase at night and pore through its contents, trying to figure out what the elder McNealy actually did. On Saturdays Scott liked to tag along to the plant and snoop around while Dad caught up on paperwork.

Like Gates, McNealy got the best private education money can buy. When Bill McNealy was tapped for a senior management job at American Motors headquarters in Detroit (he eventually rose to vice chairman), the family settled in bucolic Bloomfield Hills, Mich. Scott attended the Cranbrook School, an elite academy on a campus designed by Eero Saarinen. A jock's jock, Scott threw himself into sports, paying just enough attention to academics to get mostly B's. (He still plays hockey in a geezer league--on two teams.) Only during his junior year, when he scored a perfect 800 in math on the SAT college test, did it dawn on McNealy that he might have some academic potential. His grades picked up, and he managed to get accepted at Harvard, his father's alma mater.

Thinking for some reason that he wanted to be a doctor, McNealy signed up for pre-med courses upon arriving in Cambridge in 1972. It didn't take long for Bill Raduchel, McNealy's freshman economics teacher, to persuade him to switch to economics. (McNealy overlapped with Gates, who enrolled at Harvard a year later only to drop out in 1975, but the two don't recall meeting. "At least I graduated," McNealy chortles.) He wrote a precociously libertarian senior thesis on an ill-fated U.S. government intervention in bus manufacturing in the 1950s. But studying was mostly a sideline; McNealy likes to say he majored in beer and golf. He captained Harvard's golf team and missed the cut for the 1976 NCAA championship by a single stroke.

After Harvard, McNealy set out to seek his fortune in the factory. He landed a job as a foreman at a Rockwell International plant in Centralia, Ill., that made body panels for semi tractors. Thus began a whirlwind two-year career in Rustbelt manufacturing management, during which McNealy ran an assembly line staffed by strike breakers, served as a reliability engineer fielding automakers' complaints, and traveled the region as a factory sales rep. In his spare time he masterminded a line of molded plastic saddlebags for motorcycles. That product enabled Rockwell to reactivate a mothballed plant.

Stanford University's business school rejected McNealy twice during those years; only on his third try did his work record win out over his mediocre Harvard transcript. He enrolled in 1978, one of only a handful of students with any interest in manufacturing. The microprocessor revolution was aborning, but while many of his classmates jumped at the chance to launch a Digital Age business, McNealy dreamed only of getting back into a plant. When he graduated in 1980, he signed on as a trainee for FMC Corp. It assigned him to a factory in Silicon Valley where it was building Bradley fighting vehicles for the U.S. Army.

You could say it was fate that kept Scott McNealy in the cradle of the computer revolution, but in fact it was golf and a girl. Delighted with Silicon Valley fairways, where he could play year-round, and sweet on a secretary at FMC, young McNealy settled into the military vehicle business. (The romance didn't pan out.)

Meanwhile, Apple Computer and other startups were beginning to shake corporate America with their machines. Showing better instincts than his student for the real action, McNealy's Harvard mentor Raduchel had moved to California and was doing business with a fledgling workstation company called Onyx. In 1981, Onyx ran into production delays and quality problems; it needed a manufacturing manager. Raduchel asked McNealy for help. "I didn't even know what a disk drive was, but I could see the actual process of manufacturing a computer wasn't all that complex," recalls McNealy.

McNealy had barely learned the ropes at Onyx before a call came from Vinod Khosla, a Stanford classmate. He and some pals from Stanford wanted to start a computer company, and they asked McNealy to put them in touch with a venture capitalist who had been Onyx's CEO. Their business plan called for commercializing a workstation that an electrical engineering whiz named Andy Bechtolsheim had built for Stanford's computer network. They called their box the SUN--for Stanford University Network--workstation. The investor was intrigued; within a month, Sun Microsystems was born.

McNealy, who started out as vice president for manufacturing and operations, was hardly a front-runner for CEO. Computers weren't his thing; he needed a night course in basic electronics just so he could communicate with his co-founders. When Khosla quit to become a venture capitalist, the board was reluctant to give McNealy the top job, even though he had by far the most real-world business experience. They named him CEO only after a search turned up nobody better.

It was the smartest move the directors could have made. Like many promising startups, Sun nearly strangled on its success. The company went public in 1986 and by 1988 zoomed to annual sales of $1 billion. But its array of products grew so broad so fast that Sun began having trouble churning them out. Then it outgrew the minicomputer it used to manage production, and the IBM mainframe it hurriedly installed worked poorly at first. The factories practically ground to a halt; Sun posted its first-ever quarterly loss.

With the computer on the fritz, McNealy knew there was no hope of solving problems from the executive suite. He moved his office to the floor of Sun's biggest factory and revamped the company's manufacturing.

That much you'd expect of a factory rat. But in the months after production was rolling again, McNealy showed skills nobody expected. He ruthlessly pruned the product line, sharpening Sun's focus to workstations built around a high-powered processor of its own design, the Sparc chip. And realizing that in a well-run company of Sun's size, fixing problems on the factory floor was no job for the CEO, McNealy launched a reorganization. It pushed profit-and-loss responsibility down to individual product organizations called "planets." After that, managers felt the heat whenever things went wrong, but the results were hard to argue with. In 1992, Sun passed the $3-billion-a-year mark.

THE JOCK

McNealy does business the way he plays hockey. Although he's not that fast, he's got great lateral moves and gets in people's faces. He's great at stopping people coming down the ice, and he's also good at faking them out and getting through little holes. He gets ugly goals a lot of the time. He usually ends up with more snow on him than most people. --Curt Wozniak, CEO of Electroglas and former Sun executive

Cross an MBA jock with a factory rat and what do you get-- Not your typical Silicon Valley techno-visionary. You get McNealy: an in-your-face playmaker who knows his business fundamentals, has good moves, always keeps score--and who, when frustrated, brawls.

In most businesses, a shakeup like the one McNealy enforced in 1990 and 1991 would be enough to pull a company out of the ditch and get it racing. But the computer industry was at a turning point, and it wasn't turning Sun's way. Despite the growing demand for network servers, Sun and its Unix rivals were shipping fewer than a million units each year. The Wintel alliance, meanwhile, had sales approaching ten million PCs a year and was beginning to cast its shadow far beyond PCs.

McNealy realized that Sun's ultimate adversary was Bill Gates, not other workstation makers. Gates was touting a future version of Windows called NT, for "New Technology," that would match much of the power of Unix. NT didn't even exist yet, but Gates, with his growing celebrity, had media access to say anything he wanted. The rules had changed: Sun would need more than good execution and technical superiority. To thwart the perception that his company was a dinosaur, McNealy had to create some buzz.

He began casting about for a compelling message, a quest that did little to improve his reputation. He'd been known for high-sticking hubris since the late 1980s, when he tried to badger IBM, HP, and others into accepting Sun's version of Unix as the industry standard.

Now he directed his trash talk at Microsoft, ridiculing its products as overly complex, unreliable "hairballs" and likening them to Trabants, the sputtering autos once made in East Germany. He trotted out a slogan--"The network is the computer"--to try to diminish the perceived importance of PCs, but it only made people scratch their heads. He got much the same reaction when he named his dog Network and made it the company mascot.

He played ugly too. In 1995, after Microsoft blitzed the world with the launch of Windows 95, he surreptitiously helped bankroll the efforts of a cabal of Silicon Valley companies to prod the federal government into taking antitrust action against Microsoft. The effort came up mostly empty. For all his noise, McNealy had failed utterly to slow the Wintel juggernaut.

JAVAMAN

I remember when we turned the reins over to Scott, I was worried about how strategic he would be. Here was a guy who got his start in Silicon Valley at an Army tank factory. To hear him tell it, he'd spent his college days drinking beer and playing golf. So I told him I was worried. His eyes got small, and he said, "Just watch." --John Doerr, venture capitalist

McNealy makes no pretense of being an intellectual or a visionary. He can't remember the last time he read a book, much less a novel, and he loves to ridicule highfalutin notions about what computers can do for humanity. He isn't the kind of guy you'd expect to devise startlingly original strategies that make the rest of his industry drop everything and pay attention.

And to hear McNealy tell it, most of the high-concept strategies that he's actually hatched are borrowed from other, more mundane industries. Says he: "Our whole concept of the computer as a network device is grounded in a business model that was stolen from every other large utility on the planet. You don't have a power-generating plant in your home; you're connected to a power grid."

The model of computing as a utility network--most techies call it the "client-server model"--served Sun well in the early 1990s as corporations began to take data networking seriously. The rise of the Internet after 1994 supercharged the business, because Sun's computers make ideal Web servers.

But neither phenomenon was enough to forestall the prevailing notion that the incredible, inexorable Wintel empire would eventually swallow Sun's larger network servers too. To thwart Gates' grand designs and survive the decade, McNealy knew he needed a much more dramatic strategy--a downright visionary one that could redefine how people thought of computers. To cast his dilemma in comic-book terms, he needed a magic ring.

It turned out that he had the makings of one right under his nose. Since 1990, a team of crack programmers had been working on a new kind of programming language--code-named Oak--that would enable all manner of computerized devices to run simple programs distributed to them over a network. At one point Oak was part of an effort to develop a two-way interactive cable TV system. (Sun offered it to Time Warner for its ill-fated Full Service Network field test in Orlando, but lost out to Silicon Graphics.)

By late 1994, Oak seemed to be going nowhere, and the programmers were demoralized. Many, including their leader, James Gosling, were ready to quit and move on to one of the hot Internet-related startup companies that were sprouting all over Silicon Valley. Eric Schmidt, then Sun's chief technical officer, dreaded losing such a talented bunch, and told them to pull together one last presentation of their work, putting special emphasis on how it might be used to exploit the burgeoning Internet in interesting new ways. Do that, and he'd take it to Scott.

Recalls Schmidt: "We showed Scott the technology and described how we might want to work with partners to give it broader appeal. He saw it as much more, as a destination for the whole company, and even for the whole industry to pursue. It was like a switch went on. The moment he could map it to his problem--namely, how to harness the Internet to stop Microsoft from swallowing us all--he became its biggest supporter."

McNealy had found his magic ring--a blob of computer code soon to be renamed Java. Thus, too, Javaman was born. Within months--in May 1995--McNealy unveiled Java to the public. The concept was so novel that nobody in the computer or business press knew what to make of it.

Meanwhile, McNealy sent emissaries to show Java to Netscape, the upstart company whose Navigator browser had unleashed the Internet phenomenon. Gosling demonstrated how Java applets, or miniprograms, could add jazzy motion to pages on the World Wide Web and described how full-blown Java applications on servers could be used to handle business tasks. The Netscape guys liked what they heard and took out a license. By fall millions of copies of the first rough cut were landing on computers all over the world when users downloaded Navigator.

The more they showed Java to other companies, the more the folks at Sun knew they were on to something really big. Java held out the promise that a code cutter could write a program once, and it would theoretically run without modification on any kind of computer, regardless of the underlying microprocessor or architecture, as long as the machine was equipped with a Java interpreter program. IBM and Oracle became big supporters, as did dozens of smaller outfits. Kleiner Perkins, Silicon Valley's premier venture capital firm and one of Sun's original backers, hastily pulled together a $100 million fund to invest solely in Java startups.

With typical insouciance, McNealy claims that the Java strategy was obvious and as old as the hills: "The concept that every computer should be able to speak and understand one universal language that nobody owned was stolen from cavemen who wrote on walls."

There was just one hitch: Most of the computers in the world are Wintel PCs. To guarantee Java's ubiquity, Java had to somehow be present on most of them too. The easiest way was to embed a Java interpreter in a program, say, a browser like Netscape Navigator. Another, better way would be to persuade Microsoft to incorporate a Java interpreter right into the Windows operating system.

According to Sun cofounder Bill Joy, it was McNealy's idea that Sun pitch Java to Microsoft. "We all thought that was heresy. We also thought Microsoft wouldn't touch it. But Scott was right," he says. That's because, for once, McNealy had Microsoft's number. By the time the companies started dickering, the software giant had no choice but to license Java. In December 1995, Microsoft had decided to go after the Internet browser market pioneered and dominated by Netscape, and the only way to succeed was to match Navigator feature for feature. Java was a key feature. Microsoft paid an undisclosed sum for the license.

So suddenly Sun and Microsoft were on the same side, right-- Wrong. Constitutionally unable to simply adopt the technology of another company, Microsoft has tinkered with Java for two years now, even as Gates and his lieutenants pooh-pooh it at every opportunity. They call it merely "a moderately interesting programming language" so often that the phrase has practically become a Microsoft mantra. They argue that it doesn't make sense for important programs to be able to run on all computers because such software couldn't take advantage of unique capabilities of specific machines.

At the same time, the programmers Gates assigned to Java have busied themselves "extending" it in ways that will benefit only users of Wintel machines. If they can establish a Windows-only variant as a competing standard, it would block Sun from creating a uniform Java that can run equally well on any type of computer. (Sun claims that Microsoft, through its licensing agreement, is contractually obligated to support whatever Sun decrees to constitute Java.)

McNealy has countered by raising his sights. He's pitching Java now as a "platform" for full-blown business applications--not just dancing Web pages or word processors, but programs that manage business processes across an entire enterprise. Such applications would let companies wean themselves from Windows. As a hardware guy, McNealy became a chief proponent of network computers, stripped-down desktop computers that rely on servers to feed them Java programs so users can do their work.

Javaman being Javaman, even that wasn't enough. McNealy noisily and gleefully banned the use on Sun premises of Microsoft PowerPoint, an application that lets users create fancy slide presentations. He claimed it wasted employees' time. He also stepped up his public attacks on his nemesis, punctuating speeches with quips like, "In a world without fences, who needs Gates--"

Gates professes not to want to dignify McNealy's wisecracks with a response, other than to say: "Scott calls himself a quote machine, and he is. But you'll notice his quotes are not about what Sun does well but instead ridicule PCs and all the successful companies that have grown up around PCs."

So shrill has the rivalry between the two companies become that Aaron Goldberg, of Computer Intelligence in La Jolla, Cal., calls it a "urinary Olympics." The question is, who's got whom playing the other's game--

If there's a lesson to draw from Microsoft's growing hegemony, it is that anyone who thinks he can take the companydown, even when armed with better technology, is probably foolish. IBM couldn't, Apple couldn't, Netscape won't, and McNealy probably can't. Microsoft is just too well entrenched, too nimble in parrying threats, and too rich. The best one can hope for is to nudge Microsoft in a new direction and then try to make hay while the behemoth shifts.

By the same token, it's smart to be perceived as having a shot at derailing Gates. Customers always like to apply pressure to the big guy in hopes of driving down prices. And there's a reservoir of willingness and money in the industry to aid anyone who might loosen Gates' control over the way things will be. That's why there's so much support for Java--more than even McNealy expected.

In that sense, the game is going his way. He has forced Microsoft to rethink basic aspects of its business and play a little catch-up. When Gates counterpunches, he seems to be playing by Scott's rules. So McNealy isn't about to back down.

But here's the flip side: When you get past the rhetoric, you can see that Gates is really playing a game of attrition. That's how he's always brought his weight and wiles into play. To slow Scott down further, he'll do whatever he can to convolute Sun's efforts to get Java endorsed by industry standards bodies, his army of lawyers will reinterpret and challenge the Java licensing agreement, and his minions will bad-mouth the technology to anyone who will listen. Meanwhile, his legions of programmers will churn out one new product after another incorporating more and more Java-like features, until, in fact, it does become just another moderately interesting programming language.

Will this be curtains for Javaman-- Will nothing stand between the Evil Empire and global dominion-- Only time will tell.

Back to top

One Huge Computer

Wired Magazine
August 1998
Kevin Kelly and Spencer Reiss (also included is a conversation with Bill Joy)

Summary: The Net made it possible. Java made it doable. Jini might just make it happen. An on-the-fly, plug-and-work, global nervous system that connects his cam to her RAM to your PDA.

The Irresistible Dream: Ever since Marshall McLuhan, a central dream of the digital culture has been to create one huge computer. Not a towering superbrain tended by white-coated priests, but a vast constellation of interacting machines - processors, memory modules, disk drives, and a million other devices, all networked into a vast planetary system. A means of thinking, creating, and communicating that is everywhere at once, but nowhere in particular. A computer that is always on. Such a system would continuously spread itself and thicken, expanding by its own internal logic. It would be supremely adaptable, and hard to break. It would have myriad access points, but no CPU, no single point of failure. The global village, to coin a phrase, made real.

Engineers have long had a word for systems whose powers are widely dispersed: distributed. Banking, telephones, the electric power grid - the bigger something is, the more likely that it will be distributed. The Internet is arguably the biggest distributed system ever built, and the most complex. But all these are specialized, essentially one-dimensional undertakings - processing money, electricity, or communications bits. They pale against the ambitions of a system that aspires to be everything - to everyone.

For the biggest of thinkers, that sets up an irresistible dream: to build the network that makes all networks one, a global nervous system. The napkin sketch is simple: Take all the intelligent machines in the world - from giant mainframes to the tiniest embedded chip - and hook them together in a single intelligent network. A system open to novelty, new members, and features. A system that can tolerate what engineers ruefully call faults. A system with no limits on how large it can get, nor how small its smallest part can be.

Add a few more stipulations. To have any chance of working, the global network's structure will need to unfold from simple principles, rather than from ever more complex planning and central control. And, like another well-known distributed-computing device - the human brain - it will need to be able endlessly to reconfigure itself, to solve unanticipated problems and address unforeseeable new needs.

The key pieces for such a system - millions and billions of microprocessors - are already here, or coming. So, too, are the riotously expanding networks. Indeed, to start building that one great computer, only a single essential ingredient is missing: an architecture, a universal language, a set of superprotocols, something - and very possibly today's lexicon can't name it - to hold it all together and let the magic work. A constitution, if you like, a digital equivalent of the genetic code that all living things share.

Or, just maybe, this: a crash effort cooked up by some of the most ambitious minds ever to flee the corporate confines of Silicon Valley - a secret project spearheaded by Bill Joy, the software luminary who put the Internet on Unix and Java on close to 100 million desktops and whose fondest wish now is to give the world, to use a favorite Joy phrase, one more good "technological dislocation." He's sure he's found one. And appropriately, it's called Jini, loosely from the Arabic for magician.

General magic

In a windowless second-floor room in a deliberately obscure Sun Microsystems outpost in Sunnyvale, California, half a dozen anonymous chunks of expensive-looking hardware sit on long folding tables. Some barely rate a first look: a not particularly recent printer, what look like a pair of flat-screen monitors, a video camera, a couple of keyboards. Others are clearly prototypes: overdesigned purple computer-somethings with curved sides and stylized vents. Any Demo, Silicon Valley, USA.

Turn any of the devices around, however, and only two wires are visible: electric power and an RJ-45 Ethernet connection. Each box - even the display screens and the little handheld camera - is a fully independent network citizen, able to hold its own on the system, unencumbered by specialized cables, software drivers, or the rest of the usual array of digital life support.

Say you want to use the camera. Plug it in, and poof - a second later, an icon appears on your display screen. All the configuration chores are done automatically by one of those purple boxes - a low-end server called a lookup device - and by a 25K communication program in the camera. What's in the viewfinder-- Bring the camera image up on a monitor - any one you like. Store a clip-- The 10-gig storage device - a slightly smartened-up disk drive - is waiting. Edit-- There's another of those purple boxes, the computing device, with full workstation power. Pull some video-edit software out of the storage module, and you're off.

That's one possibility. Or maybe you'd rather batch print some letters from your laptop. Done. Or get that old laser printer online. A pocket-sized adapter does it. Or add another 10 gigs of storage - no need to call a sysadmin, just grab a drive off the shelf, and plug it in.

On one level, the demo is the ultimate in plug-and-play technology - "plug-and-work," its Sun-shirted minders note with a smile. No mean feat. Not surprisingly, some of the Jini demo's most interested visitors have been from hardware companies that would dearly love to find a way for us all to snap a few billion more microprocessors, disk drives, and other smart devices into our personal networks.

But Jini aims much higher. What Joy and the two dozen programmers working with him aspire to do is nothing less than dynamite the whole creaky logjam of computing, as it has evolved from giant mainframes through the first clunky PCs to today's cobbled-together Internet and Windows Everywhere. If they succeed, Jini code will provide connections that will make today's information "superhighways" look as confining as 19th-century railways. And that, Jini thinking goes, will be the foundation for truly networked, global computing - organic and ever changing, and keyed to a hurtling future instead of being shackled to the platforms and conventions of the past. "When the foundations are so far off," reads an internal Sun document written to support the project last year, "it makes sense to do a reset."

Coming from almost anywhere else, that declaration would be laughable. But Sun and Bill Joy have come close once already to pushing computing's reset button, with its still-expanding programming language Java, the most important development in computing since the explosion of the Internet. What Java aims to do for software - be a lingua franca - Jini hopes to do for the machines that run it: provide an overarching, universal platform - a distributed operating system, in effect, on which devices of every description can meet. "Jini is the next chapter in the Java story," reads another project mantra.

And Jini is no clunky hack, strung together in a lab with glue and wire to impress the boss and calm investors. Most of the demo devices are modified versions of existing hardware - one of the project's driving ideas is to not have to throw existing systems away. Jini software has been in limited-release beta since June, with testing under way by some of the biggest names in computers and consumer electronics - NEC, Toshiba, Quantum, Ericsson, Siemens, Computer Associates, and a dozen others. By the end of the year, Sun hopes to release a full package, from a network infrastructure to the little 25K program that can put your front-door light switch onto the network. The release name is still being debated, but the marketing plan is not: It will reprise the same strategy that fueled the explosive take-offs of both the World Wide Web and Java - essentially, give it away. "There's one thing we've all learned from watching Java and the Net," says Mike Clary, Joy's key colleague in Aspen and Jini's overall project manager. "This can only be a ubiquity play."

Jini's prelaunch team shares a building with what remains of another audacious attempt at networking heroics, General Magic - a reminder of the casualty rate of would-be technological revolutionaries. A Jini victory would mean the creation of a loosely connected federation of computers freed from today's OS tyrannies - one reason not to expect a friendly Microsoft embrace. Neither Bill - Gates or Joy - needs reminding that it was the modest little PC's universal appeal, not the US Justice Department, that ultimately humbled IBM's mighty mainframes. And if lightning strikes again, those anonymous boxes in the windowless demo room could someday end up in a technology museum: cell zero of the global computer. Not to mention giant slayers.

If ...

Up from Java

Bill Joy doesn't like the word "exile," but he's made a second career out of keeping most of the Rocky Mountains between himself and Silicon Valley. A founder of Sun Microsystems and still officially Sun's VP for research, Joy took himself to Aspen a decade ago to build a geek-lord's dream: his own custom research-and-development lab, aka Sun Aspen Smallworks. Small-- "Ideas resemble the organism that built them," Joy says, "so a small organization will build simple things that work." Meaning what-- "The idea is that we do whatever is most important - not necessarily most urgent. Sun has 20,000 other people doing that. I left the urgent behind to get to the important."

In the early days, Joy and a rotating Smallworks crew focused on what they dubbed the "4MY" program - "Four Miracles a Year," everything from microchip design to networking theory. More recently, Aspen was a refuge for the long-running project that became Java, from its early near-death experiences as "Green" and "Oak" to the first big licensing deals.

It's a pleasant place, Smallworks, behind one of those too-cute Victorians above the year-round commotion of shops and restaurants in downtown Aspen. Joy and a couple of permanent staffers inhabit a cheerful clutter of exotic gear, whiteboard, and piles of books. But blissed-out the view definitely ain't - the view of the high tech landscape, anyway. "We're in the Dark Ages," Joy says, wheeling out his favorite rant. "It's 900 AD - medieval computing. Except for the Web, what's really getting better-- I managed to get my notebook computer to talk to the printer - it took a month. Our basic operating systems now have some 20 million lines of code, and more is being piled on every day. It's insane to try to build the future on that."

Indeed, from whichever angle you look - Silicon Valley prince or baffled user - complexity and scale are the mad aunts in the attic of today's computing. Lines of code piling up like crust on hard drives are only part of the problem - the real nightmare starts when you add blossoming networks to the mix. Systems engineers measure complexity with a metric: number of users times number of machines times number of functions being undertaken. Put a couple of those numbers into the millions or billions - which the Net explosion is doing - and you get unmanageably huge, quickly. Unless, of course, you have a system that can pull order from networking chaos.

Visionaries and hard-headed engineers - not to mention Windows-for-all Gates - have been groping for years to find paths through the spreading complexity. Ted Nelson's Xanadu, Xerox PARC's Smalltalk, David Gelernter's Linda; the list is long and not encouraging. One general path has been idealized - start-from-scratch systems, most of them quixotic or mainly research ventures. Another, less sweeping approach has been object-oriented programming - building applications on the fly from small code modules, usually called objects, the better to move them around a network or translate across platforms. Two rival object standards, the industrywide Corba and Microsoft's DCOM, have kept sprawling corporate networks from degenerating into towers of Babel. And then, of course, there's the one unalloyed success story of distributed computing: the Internet, and its prodigal, the Web. Ironically, though, TCP/IP's very success in creating a global medium has only made the overall problem of complexity even worse.

As the Net's explosion gained force three years ago, Bill Joy was deep into Sun's own object-oriented programming effort. The motives for releasing Java - an elegantly stripped-down language originally designed to run consumer electronics - were less than pure and more than a little desperate: to blunt "Windows Everywhere!" with a new technology that promised platform independence. The ability to run the same program on any computer - Mac, Windows, Unix, a tiny device on your wrist - is a key distributed-computing tenet, not to mention an obvious boon to a global information network. Skeptics laughed nonetheless. But the timing was perfect - even more so when Netscape, looking for allies and ammunition against Microsoft's gathering counterattack, built Java compatibility into its runaway-hit browser. What might have been another high-minded experiment instead became an instant global standard.

Sun from the start has famously been the company that preached "the network is the computer." But even for Joy, holed up in Aspen writing the Java specs, that explosion was astonishing. Though Java was launched as a new programming language, Joy and the others had always assumed that they would slowly build it into a full software platform - one that really fulfilled the brash early promise of "write once, run anywhere." Their best guess had been that it would take five years to achieve what they reckoned was the critical mass needed to launch a viable distributed platform - about 100 million users. But the Net's amazing growth had them scrambling almost immediately. The good news was that Joy and the rest of Sun's software research team already had a clear sense of where they wanted to go. "We knew that whatever we did had to be technically simple," says Joy, "because it's hard to write programs, and even harder to write distributed programs - you have the whole big complicated system to think about. What we wanted was a very simple communications mechanism that would let the distributed system work."

One of Joy's favorite engineering maxims - "Large successful systems start as small successful systems" - is another way of saying: Use what already works. In 1994, the Aspen skunkworks already had a workstation running Oberon, an ambitious attempt by Z�rich-based Niklaus Wirth, the inventor of Pascal, to create a featherweight system written entirely in one simple programming language. Such knowledge-based computing erases the conventional distinction between the OS and applications. Building distributed networks, Joy believed, was a key breakthrough. Another intriguing model was Gelernter's Linda, whose central idea, called "tuple spaces," is a radically simple way to organize communication between software objects; Linda's broad concepts had already been adopted for JavaSpaces, a tool for building distributed applications.

And then there was Java itself, which continued to build momentum among programmers - and with that, more and more of the plug-and-play software components crucial to making object-based programming work.

In the spring of last year, Joy sat down in Aspen with Sun senior staff engineer Jim Waldo, whose research group had just completed Java RMI - Remote Method Invocation, an interface tool that lets distributed software objects find and communicate with each other over a network. Sketching on - yes - a napkin, they realized that the practical outlines for a full-out distributed-computing system were already visible. They also had the people, based mainly in Sun's East Coast software research lab in Chelmsford, Massachusetts. Waldo himself had already started the basic code for what programmers call transactions, which ensure that groups of commands sent out over the network actually occur as a unit. A variety of programmers had worked out leasing, a framework for short-term relations between objects. Bob Scheifler, a leader of the X Consortium - an industrywide initiative to build cross-platform interface technology - had the network-security know-how. "Two coffees into breakfast," recalls Joy, Jini was in high gear.

Reality check

You have to drill down energetically into Microsoft's sprawling Web site, but there it is, in the list of projects under way at Microsoft Research. "We believe it is time to reexamine the operating system's role in computing," reads the opening line of a proposal for an initiative dubbed Millennium, described as "a new self-organizing, self-tuning distributed system." NT's 20 million-odd lines of code notwithstanding, Millennium envisions "a distributed operating system, based on a few principles pervasively applied." As part of that system, "any code fragment might run anywhere, any data object might live anywhere." Sound familiar-- It would also be "self-configuring, self-monitoring, and self-tuning. And of course, it would be scalable and secure." Of course.

In the long tradition of Microsoft vaporware, there may be less to Millennium than meets the eye - the team consists of a half-dozen full-time researchers, according to a spokesperson, and a couple of active prototypes. One working system, dubbed Coign, distributes conventionally written applications on the fly; the other, Borg, creates a distributed version of the Java Virtual Machine. Microsoft famously got jumped once before by a technology, the Internet, that didn't quite fit Redmond's worldview; despite all of his current distractions, Bill Gates doesn't want to get paradigm-shifted once again. Whatever Millennium turns out to be - vaporware stalking-horse or shrewdly hedged bet - the Kremlin of centrally planned computing has more reasons than most to be paying attention to new rumblings on the network.

And there's not just Bill Joy and Sun to worry about. Add to the list Lucent's ever about-to-take-off Inferno; an ambitious Caltech project called Infospheres; even Larry Ellison's half-baked network computer scheme - all are pursuing the distributed-computing dream. Even sleepy AT&T this spring unveiled a Java-based "enhanced network infrastructure" called GeoPlex, designed to let telecom companies offer services across the whole array of digital devices and networks. Apparently you don't need to be a software hero with a private Aspen research lab and 20/20 programming vision to detect a potential revolution.

So ... why Jini--

The short answer, of course, is Java, whose slipstream - a million active programmers, by Sun's latest reckoning - can give Jini the kind of instant presence and easy learning curve Java got from Netscape and C++. A quiet argument is under way in Jini's marketing team over how closely to stick to Java branding; the leading contender has been JavaTone, as in the universal telephone signal.

But Jini is primed to ride a potentially even more powerful new wave: hardware geared for the network. Jini's main beta testers are not the usual Silicon Valley coders - licensees already signed up include a dream team of big-time hardware players who seem to be falling over each other with raves. "Anyone who's ever tried adding storage on a LAN can tell you why we need this," says Paul Borrill, vice president and chief architect at Quantum, the disk drive maker. "To use an overused phrase, this is a paradigm shift." Quantum expects to ship its first Jini-ized devices late next year. Billy Moon, Ericsson's New Concepts program director, goes one shift better: "It's a double-barreled paradigm shift that reaches beyond the computer industry. The combination of componentized software running on distributed virtual machines and the bold system architecture transform and blur the very idea of what computers, networks, and applications are."

Things get vaguer when the question turns to the new services that Jini could spawn. Plug-and-play is a nice feature. Exploding the computer back to its components - storage and processing especially - is a potential revolution, opening the door to everything from supercomputing on demand to massively encrypted remote data storage and your own personal desktop available on any machine in the world. Clever corporate marketers and ecommerce entrepreneurs presumably will sort these offerings out.

Jini avoids one common stumbling block of many clean-slate solutions: incompatibility. Specialized programming languages, legacy applications, and hardware all do fine under a Jini r�gime; the only requirement, beyond being Java-enabled, is that they observe the basic networking rules. "The whole idea is to be very forgiving," Joy says. "If you have slightly different code than I do, that's fine - when I get one of your objects, I also get the code that goes along with it. We don't have to agree beyond the basic rules, and we can let the best - the most functional, fastest, easiest - code win. So you can keep your Windows if you want it. But now the network will be evolutionary - the survival of the fittest."

But of course, "fittest" in technology does not always mean "best" - hello, Macintosh and Betamax. On the Net and in court, Sun is already battling competing Java "flavors" - variations of the language - launched from Redmond. In May, Sun filed a suit against Microsoft to try to rein in its licensees and enforce "100 Percent Pure Java." But the fight has at least given Jini's creators the benefit of hindsight. And the sidestep they came up with plays directly to the strength of a distributed system: When Jini tries to run on a nonstandard Java Virtual Machine, Jini automatically queries its capabilities, then uploads whatever chunks of code are needed to make it fully compatible. "You could design a system to prohibit that," says Clary, the Jini project manager. "But that would violate the licensing terms a lot more flagrantly than just leaving some features out. There'd be nothing gray-area about it. And it's hard to see the value in deliberately shutting yourself off from the world."

Sun has also been working overtime to address what remains the favorite bugaboo of Java skeptics: speed, as in lack thereof. Java's "sluggishness" is a favorite complaint of Net surfers watching Web applets - Java's most visible face - slowly unfurl. The seriousness of the problem has attracted a correspondingly high amount of programmer energy. As one result, a new generation of just-in-time compilers is emerging for a variety of operating platforms, produced both by Sun itself and by third-party developers. And later this year, Sun will release the 1.2 version of Java, one of whose new features, HotSpot, is dynamic optimization, which Sun officials claim can take JIT compilation to "C-level performance."

Will Jini scale out to the size of, say, the planet-- "We've looked at this every way we can think of," says Clary. "And the answer is yes." Object-based programming makes sense for the same reason that packet-switching is now the technology of choice for networks: It reduces huge problems to small pieces. That in turn points inevitably to a move from today's mostly client/server networking to peer-to-peer relations, with code and data flying in all directions across the network. And the resulting complexity, the Jini team concluded, could be dealt with only by stripping its basic operating rules to an absolute minimum. "How do we know whether we made the right choices--" says Waldo. "You never know. We stopped only when we couldn't throw things out anymore."

When Joy and Clary took Jini to Sun CEO Scott McNealy for a green light in March of last year, they used the phrase "opportunity driven" - Valley-speak for a project that will build its market on the fly. As with Java, the benefits to Sun are a subject for debate - possibly, Jini-configured hardware; more certainly, an inside track on what could well be historic technological changes. What everyone agrees is that timing will determine Jini's fate. "It's like that portal opening in Star Trek," Joy says. "If you're lucky, you get through the opening, and then the portal closes."

Comes the comet

In 1979, Steve Jobs - then an unknown 23-year-old geek - made his now-legendary visit to Xerox PARC to see the radical new Alto computer, with its primitive mouse and icon-based screen. "I saw a very rudimentary graphical interface," he said years later (see "Steve Jobs: The Next Insanely Great Thing," Wired 4.02, page 102). "It wasn't complete. It wasn't right. But within 10 minutes it was obvious that every computer in the world would work this way someday." Two years ago, Jobs made the same prediction about object-based distributed computing. "You can argue about how many years it will take," he said, "and who the winners and losers will be during this transition. But you can't argue about the inevitability."

WebObjects, Jobs's project in pursuit of that vision, never took off in part because its success depended on a wholesale switch to a new hardware platform, the ill-fated NeXT. Not an especially good strategy for an undertaking with universal aspirations.

But, as Jobs predicted, one way or another it will happen - indeed, it is happening, before our very eyes. The Web is growing in every dimension - faster, bigger, deeper and more sophisticated by the day. Intelligence is being embedded in everything. Ever larger chunks of human activity are migrating to the network. And that greater genie surely will not be going back into any bottle.

Joy's Jini, if it takes hold, has the potential to overturn the familiar territory of hardware, personal computers, peripherals, phones, TVs, and appliances. The vision of what comes after is just that - a vision. For people like Bill Joy, it hovers like a city on a hill, elegant and platonic, waiting for us humans to make it so. But the closer it gets, the easier it will be for everyone to see. "Imagine a global network so complex it will be a kind of organism, a dynamic, richly interconnected medium wrapped around the earth 24,000 miles deep." That's not Teilhard de Chardin - it's the 1997 annual report from Daimler-Benz North America.

For now, though, some old lines are still drawn: central planning versus competition. NT's 20 million lines of code versus the 600 Kbytes of Jini. Bill versus Bill. Redmond versus Aspen - there's a pattern working here, and it almost surely has as much to do with philosophy or faith as it does with questions of mere technology.

For its part, Jini is gambling that a small nudge can actually relocate a mountain. "Our goal is to lose control over the network," says Jim Waldo, "and make everyone else - from Bell Labs to Redmond - lose control too." He's not talking about market share, not by itself anyway. "What we're trying to build are the mammals to compete with the big computational dinosaurs. You can imagine how the conversation went: 'They're too small. They're nothing - they're not enterprise scaled.' But the comet is coming. And when it does, we know who inherits the earth."

Jini is a set of new software layers that together create an overarching "federation" of computer devices and services.

On top is a directory service, based on a "lookup" mechanism that allows different Jini-enabled devices and applications to register and be seen on the network. The next-level service is persistence, provided by JavaSpaces technology, which stores objects so that other users or applications can retrieve them. Below that, a set of protocols based on Java's Remote Method Invocation enables objects to communicate and pass each other code. And finally a boot, join, and discover protocol allows Jini-compatible devices, users, and applications to announce themselves to the network and register in a directory.

Any device with an operating system capable of supporting a Java Virtual Machine - meaning, in practical terms, any modern computer - can be linked with a Jini network. Simpler devices can also join, though on a more limited basis.

JavaSpaces are virtual "bulletin boards" or "marketplaces" - the heart of Jini's distributed networking. Using a few simple programming methods, including "read," "write," and "take," JavaSpaces make software objects available to anyone in a network. The objects themselves can define a job to be done, a problem to be solved, or a service being offered. A JavaSpace can be as small as 10K and as large as 100 Mbytes.

Back to top

Sun's Bid to Rule the Web

Last year's eBay crashes prompted Sun Microsystems to embark on a mission: To become as dominant and reliable as Ma Bell was
July 24, 2000
BusinessWeek
Peter Burrows

Late last June, Sun Microsystems Inc. (SUNW) President Edward Zander got the kind of call every tech executive dreads. After eBay Inc. (EBAY) suffered a 22-hour outage of its Web site and a spate of smaller crashes, CEO Margaret Whitman called to tell Zander that the problem was a bug in Sun's top-of-the-line server. Sun would learn something just as startling over the next few days of round-the-clock meetings with eBay: The Internet upstart didn't have a clue about running a $1 million-plus computer. The company hadn't provided sufficient air conditioning to keep the machine cool. And even though there had been a software problem with the machine for which Sun had issued a patch many months before, eBay had simply neglected to install it. The list went on--fueling the sentiment, as one Sun manager put it, that "selling computers to some of these dot-coms is like giving a gun to a 5-year-old."

That's when Zander realized things could get much worse. For most dot-coms, starting their business on a Sun server is almost a given. Already, more than 40% of the servers found in the computing centers that house most Web sites are Sun's, and that market is expected to boom as everyone from new Net companies to the click-and-mortar crowd set up shop online."It suddenly hit me," says Zander. "How many future eBays are buying their first computer from us this very minute--" Adds Sun CEO Scott G. McNealy: "It was our Pentium moment," comparing the eBay incident to the lesson Intel Corp. (INTC) learned in 1994 after the chip giant angered customers by initially trying to downplay a bug in its new Pentium chip. "That's when we realized it wasn't eBay's fault," says McNealy. "It was our fault."

McNealy and Zander didn't need another wake-up call. Since then, the two have been tearing apart Sun and rebuilding it in an effort to make the Net as reliable as the telephone system. Just as AT&T (T) became Ma Bell, providing that always available dial tone, Sun is shooting for no less than Ma Web, the supplier of super-reliable Web tone. To do that, Sun is moving far beyond Web servers to providing many of the technologies required to make this possible: storage products, a vast array of e-business software, and consultants that not only supply all the gear but also hold customers' hands every step of the way.

Safe bet. If the duo can pull it off, Sun could emerge as the King of the Net--every bit as dominant as Big Blue in its mainframe heyday or Microsoft Corp. in the PC era. Just as high-tech managers used to say, "No one gets fired for choosing IBM (IBM)," Zander aims to have the same said of Sun. "I want to be the safe bet for companies that need the most innovative technology," he says.

Sun hopes to go down in the history books as that rare company with the vision to change an industry and the ability to cash in on that vision. Since it was founded in 1982, Sun has promoted the notion that "the network is the computer," a view of computing where the action isn't on desktop PCs but on big central servers where computing can be doled out in easy-to-use chunks, wherever and whenever desired. With the explosion of the Internet and rapid deployment of high-bandwidth networks, Sun's vision finally is becoming a reality. "McNealy held out for the pot of gold," says Bill Raduchel, a former Sun executive who is now chief technologist at America Online Inc. "It took a decade to play out, but now the pot of gold is here."

That's why Sun has been on a tear. In the most recent quarter, revenue climbed 35%--more than any other computer company, including PC darling Dell Computer Corp. (DELL), which grew 30%. Sun is growing faster than at any time since 1991, when it was one-fifth the size it is today. And with gross profit margins of 52%, it is the most profitable computer maker in all of techdom.

McNealy vows this is just the beginning. Known for having the strategic vision, slickest sales reps, and hottest new products--but not the best service--Sun has made reliability the top priority. That means pumping up the services business and overhauling the way the company designs and sells its products. In the past year, Sun has reduced the number of configurations it sells from thousands by pushing customers to choose from under 200 models. And now, managers and sales reps are compensated largely on customer satisfaction. What's more, McNealy, a sometime golfing buddy of General Electric Co. (GE) Chairman John F. Welch, has become a convert of GE's Six Sigma quality program that builds in checks to make sure customers' operations stay up and running. By far, the boldest element of McNealy's plan is software. Sun is trying to define and dominate a new category of software that combines many of today's e-business software segments, including e-mail, e-commerce portals, and programs for serving up Web pages and wireless applications. The idea: Wrap a suite of applications into one fail-safe whole available on any Sun server. On July 17, iPlanet, Sun's Net software joint venture formed with AOL last year, unveiled the new suite, along with an audacious goal: Within 18 months, the company expects to hit the $1 billion mark in e-commerce software sales, according to Margaret Breya, iPlanet's vice-president of marketing. By 2005, she says Sun could havea $5 billion to $10 billion software business. Other executives, however, say it may take a buying binge to get there.

Put it all together, and Sun is designing its own take on an old trend: vertical integration, in which it sells software, hardware, and services as one--just like telecom equipment makers Lucent Technologies (LU) or Nortel Networks Corp. (NT) do with their phone switches. "The computing model of tomorrow is the telecom model of today," says Masood Jabbar, Sun's senior vice-president of sales. How does Sun fit in-- It plans to make the "big frigging Webtone switches," as McNealy calls them--the powerful servers that can whisk billions of bits around the Net, along with the software that manages Web pages, dishes up data, and executes transactions. "The world's moving in our direction at 8 gazillion miles per hour. Our biggest problem is just trying to keep up," says McNealy.

That's why he has lit a bonfire under Sun. After the eBay incident, Zander called a meeting of all managers and read them the riot act. Late last summer, his staff identified 14 key initiatives, such as new processes for conducting customer audits, with one of Zander's top vice-presidents in charge of each. And on July 1, McNealy reorganized Sun, combining fiercely independent sales operations within product units into one single sales organization. Now, customers see one sales rep for their entire business, instead of being bombarded by reps from different divisions. And McNealy has created a Customer Advocacy Organization to make sure all divisions are putting reliability and customer satisfaction first. Division president Mel Friedman, for instance, has authority to request the redesign of any Sun product for suspected glitches. Says Breya: "It's about Sun growing up."

As we all know, though, growing up is hard to do. For Sun to shake off its upstart ways, it will have to make the shift from an engineering-driven company to a full-service company. That means mastering software sales, a historic weakness, and building up consulting to help companies design their e-businesses around Sun gear. And it must do all this while holding off heavyweights such as IBM and Hewlett-Packard. The stalwarts may have been slow to grok the Net, but they have a legacy of ultra-dependable products that could be a major advantage. "Sun rode the wave of dot-coms, but those companies have different needs now. And taking care of those needs is IBM's and HP's forte, not Sun's," says Bruce L. Chovnick, senior vice-president at Network Solutions, a Web registry company that recently ditched a Sun high-end server for a mainframe from IBM.

McNealy will have to stare down other challengers, as well. At a time when servers based on Sun's new UltraSparc3 chip are a few months late, longtime PC industry rivals are massing for yet another assault on the server market. Using Microsoft's (MSFT) four-month-old Windows 2000 program or the free Linux operating system, PC makers will continue to chip away at the market for less powerful servers--especially after Intel brings out its new IA-64 chips, due by yearend. "Customers are willing to pay high prices and go with the safe bet [Sun] in these early days of the Net. But ultimately, we'll be able to redefine the economics of the Internet," says Compaq Computer Corp. CEO Michael D. Capellas. Adds International Data Corp. analyst Jean S. Bozman: "Everyone is shooting at Sun, there's no question about it."

The company with the most ammunition is Microsoft. On June 22, Microsoft announced its version of Sun's Webtone scheme--an initiative dubbed .net that is designed to make the Web much easier to use. In it, unrelated Web sites, Net services, and traditional Windows software programs can be linked together to do useful things--say, to get your bank's Web site to transfer money to your e-broker, who buys a stock and then records the trade to your Microsoft Money program on your PC. Such complexity requires software expertise, snorts Microsoft CEO Steve A. Ballmer, "and Sun'snot really a software company." Counters Sun chief scientist Bill Joy: "I've been writing about network-based computing for 20 years. Microsoft embraced it last week."

Sniping aside, Sun faces even more software challenges. Throw into the mix programs such as Napster that make it easy to link files directly from PC to PC, altogether bypassing huge servers, and some analysts think McNealy & Co. could face a resurgence of powerful PCs that can store and move data around the Net. That could put a squeeze on server profits. Sanford C. Bernstein & Co. analyst Toni Sacconaghi thinks profit margins for Sun's servers could fall from the mid-50s to the low-30s within three years. So it's crucial that Sun crank up sales of hugely profitable software and storage products, with gross profit margins of 80% and 60%, respectively.

Only then can Sun continue to fund its $2 billion research-and-development effort and keep spending at an industry-leading rate of 10% of revenue. If it can't, Sun may find itself boxed into a high-end corner of the computer industry, adding to the list of once proud computer companies such as Digital Equipment Corp. that have been whittled away by PC makers.

Sun has managed to outfox the doomsayers before. In the early 1990s, when profits collapsed for the technical workstations that brought in 90% of the company's revenue, McNealy bet the next big opportunity would be servers. He poured billions into developing technologies such as the Solaris operating system. Now, servers and related gear bring in roughly 80% of Sun's $11.7 billion in sales. Even more remarkable is Sun's assault on the high-end server market once dominated by IBM mainframes. While the market for $1 million-plus servers shrank 17.8% last year, to $11.4 billion, Sun's revenue has rocketed 28% because of runaway sales of its e10,000 Starfire machine, according to IDC.

Unlike high-tech dynasties such as IBM or Microsoft, Sun's grand plan is not based on locking customers into its own proprietary technology. IBM and Microsoft modulated the flow of new technology in the mainframe and PC eras largely by maintaining control of technical interfaces that others would need to create compatible programs and peripherals. But Sun wants to dominate Internet-style--that is, by doing as much innovation as possible, licensing leading-edge work as the standard for others, and then racing to stay ahead.

That puts the pressure on Sun's big thinkers, like Joy (page EB 42). For starters, Joy and Sun's other technologists have coined the term "Net Effects" to describe the challenge of keeping up with spiraling demand as a billion people use the Net more often, from more devices, and in different ways over the next few years. To keep pace, Sun's servers will have to accelerate in power at a rate at least 100 times faster than Moore's Law, which holds that chips double in speed every 18 months, says Sun chief technologist Greg Popadopolous. Sun is working on two tracks--massive single machines with millions of microprocessors, as well as distributed computing schemes so the computing load can be divvied up between smaller machines linked by high-speed networks.

Sun also is betting it can leapfrog the competition by giving customers the essential software they need to run their e-businesses in one neat, foolproof package. Today, companies face a blizzard of offerings--application servers to host and handle e-mail, Web servers to manage and send out Web pages, and portal programs on which to give the sites a unique look and feel. While these stand-alone software products may deliver the latest bells and whistles, it costs a fortune in consulting fees to make them work together.

Sun's approach is different. iPlanet packs snazzy programs into a suite known as the Internet Service Deployment Platform. Don't be fooled by the clunky name. Using this suite, customers can get up and running quickly because Sun has made sure the software works in sync. With the price starting at $500,000, Sun isn't looking to undercut the competition. Instead, customers will save on installation. "This could cut my development time by 30%," says Norbert Nowicki, a senior partner with Computer Sciences Corp., an El Segundo (Calif.) computer services consultancy.

Sun isn't the only company offering such a suite. Oracle (ORCL), IBM, and Microsoft do, as well. But none of those companies is the dominant provider of the computers on which the software must run. "Sun isn't just dragging the software along anymore," says Goldman, Sachs & Co. analyst Laura Conigliaro. "It can be a serious driver of new business." Especially with partner AOL (AOL) using the software suite internally and promoting it to its Net customers. "AOL is customer No. 1 for iPlanet," says David Gang, an AOL executive who recently became iPlanet's executive vice-president. "If we can build products that satisfy AOL, it should work for everyone else."

The irony of McNealy's software approach is that he's stealing a page from the Microsoft playbook--a twist on Microsoft's "embrace and extend" strategy of absorbing fresh technologies into its Windows software. Instead, Sun wants to either bundle or weave Net software into its Solaris operating system. The process already has begun. While competition used to be fierce in the market for arcane directory software, where companies store their databases of employees, customers, and suppliers, now Sun dominates because it has embedded directory software into the latest version of Solaris. "This could be every bit as big as Oracle's [$7.4 billion database] business," says Mark Tolliver, general manager of iPlanet.

In recent months, the company has made a push into hot new areas, such as a wireless server that will go head-to-head with IBM and others, and e-commerce and e-marketplace applications that will compete with offerings from Commerce One, Oracle, and others. And while iPlanet doesn't have a product to rival red-hot programs like Vignette's software for managing Web pages, Sun may develop offerings in this niche or buy the pieces necessary to offer it. "With our stock where it is, we'd be remiss if we didn't look at this," says Jonathan Schwartz, recently named Sun vice-president for corporate strategy.

Storage Breakthrough. One area where Sun hasn't been able to get off the ground is storage. The company has made two failed attempts to introduce new products in the past three years. "This business takes focus, but storage was an afterthought for Sun," says Raduchel. No more. Sun claims it has made a breakthrough and has created a specialized sales and support organization to push it. Never mind lining up big cabinet-size storage racks tethered to servers--the way most storage farms operate. Instead, customers put Sun's new T3 storage boxes wherever makes the most sense--without having to be within close proximity to a server. An Internet service provider, for example, could put one in a Boston office to speed Red Sox scores to the locals--regardless of whether that site uses servers from Sun or a rival. "The upside for Sun in storage is immense," says Goldman's Conigliaro, who thinks Sun's $2 billion business will grow 25% a year for the next three years. Still, in that time frame, rival EMC Corp. is expected to shoot past the $15 billion mark.

When did Sun get so serious about growing up-- Rumblings began in 1998, when Sun's brain trust began to sense that customers' needs for keeping their Web sites up and running were far outstripping Sun's knowhow. But for McNealy and Zander, the eBay incidents in mid-1999 underscored how fast those requirements were rising--and far behind Sun really was.

Sun sprang into action to solve eBay's problem, and within weeks, it worked out a plan with software partners Oracle and Veritas Software Corp. to stabilize eBay's server--even devising back-up systems that have kept eBay out of the news despite six or so crashes in recent months. "We were pushing Sun's products to places they'd never had to go," says eBay Chief Technology Officer Maynard Webb, who last fall nearly switched to IBM. "For Sun to still have our business is a testament to their ability to solve those issues."

Zander was worried it was more like dumb luck. He knew last-minute heroics would not be possible should eBay-like debacles become commonplace. So in early July, Zander assigned Vice-President John C. Shoemaker to come up with a set of initiatives to meet customer demand for rock-solid gear. By the end of August, after key areas for improvement were identified, Zander decided it was time to turn up the pressure inside Sun, calling for daily 8 a.m. meetings with the management team to discuss any problems at customer sites. "Scott and I decided to ruin everyone's morning," he says.

Now, all high-end systems must be pre-tested with the customer's software before they ship. Another team is making sure that all new products can be monitored remotely from one of Sun's data centers, finally bringing it up to speed with rivals such as EMC and IBM. Sun has also done two-day, lengthy audits of 75 top customers, sometimes issuing 100-page reports that recommend making changes such as adding a humidity sensor to ensure that atmospheric conditions are optimal for Sun equipment.

And McNealy has become a crusader for the new quality program, dubbed Sun Sigma. Now, Sun's top execs will get four days of training and will then lead teams that will get four weeks of training in Six Sigma-style practices. Any manager who doesn't lead such a team over the next 18 months, says Zander, can forget getting promoted to vice-president.

Why the hardball tactic-- With 35,000 employees, Sun will have to start behaving less like a mob of high-tech freedom fighters and more like an icon of big management control. If McNealy can pull that off, then Sun might one day truly be worthy of the nickname Ma Web.

Back to top

The Odd Couple

Scott is new economy; Jack is old. Scott wings it; Jack has it all mapped out. But thanks to golf and the Internet, Sun CEO Scott McNealy and GE's Jack Welch have become best buddies who mentor each other. They talked to FORTUNE about the Net, creating great companies--and, of course, golf.
Fortune Magazine
May 1, 2000
Brent Schlender; John Huey; Scott McNealy; Jack Welch

Baby-boomers can rightfully claim credit for providing most of the digital technology and entrepreneurial imagination that unleashed the new economy. Although the building blocks of microprocessors, personal computers, servers, networks, and interactive software were gestating in the labs and office cubicles of old-line high-tech companies, it was boomers in denims and T-shirts--not their buttoned-down elders--who over the past two decades wove all these together into the ubiquitous information and communications infrastructure that supports a whole new way of doing business.

But as any number of these now-graying whiz kids will tell you, there's still a lot to learn from captains of the old economy. After all, Microsoft's Bill Gates regularly consults Warren Buffett for advice, not to mention an occasional game of bridge. Steve Jobs, the erstwhile rebel who clashed with his elders during his first tenure at Apple, often sought out the wisdom of Polaroid founder Edwin Land before he died in 1991.

Thanks to that most traditional of business sports--golf--there's now another inter-generational, old economy/new economy CEO buddy act. It pairs General Electric's Jack Welch, who at 64 is perhaps the most admired CEO of his generation, and Sun Microsystems' Scott McNealy, who at 45 seems one of the shrewdest and most forceful personalities in the computer business. Both are as competitive as they come.

At first blush, their companies don't seem to have much in common. GE is a $112-billion-a-year holding company for an array of businesses ranging from financial services to broadcasting to jet engines to, of course, light bulbs. Sun has one-tenth GE's annual revenues and basically makes one product--computers that undergird the Internet. Welch can count his competitors by the hundreds; McNealy's boil down mainly to Microsoft, IBM, Compaq, Dell, and Hewlett-Packard. Jack commands 340,000 employees; Scott oversees 37,000.

This is where golf comes in. In 1998, McNealy and Welch were ranked No. 1 and 2 in Golf Digest's listing of top CEO golfers. McNealy figured that if he challenged the GE CEO to a shootout, he could get some quality face time--at least 18 holes' worth--to pick Welch's brain about managing a huge corporation. They hit the links on Nantucket that summer. McNealy showed up with a golden trophy called the Welch Cup for the winner, and promptly got beaten. (No, he didn't throw the game. Yes, we asked.) But a genuine friendship took root, and McNealy later got a consolation prize of sorts--a seat on the GE board.

Over the past 18 months that friendship has deepened. McNealy has stoked Welch's growing interest in e-commerce, and his irreverent personality has enlivened GE board meetings. For his part, Welch has shored up McNealy's resolve to transform Sun from something of a seat-of-the-pants operation to a more disciplined organization that, like GE, is obsessed with product quality and customer service.

Early in March, McNealy and Welch sat down with this writer and FORTUNE managing editor John Huey to talk about their friendship, their companies, and where the new economy meets the old. We met following a board meeting at GE's Rockefeller Center headquarters in New York City. It was a balmy Friday afternoon, and the only one in the room who was sporting a tie was your reporter.

FORTUNE: Four years ago, Jack was standing at the podium with Intel's Andy Grove at a Fortune 500 CEO Forum in San Francisco, and he basically told everyone, "I don't have a computer in my office, and I don't need a computer." Obviously Scott comes from a different place. Eight years ago he was presciently saying that e-mail is a killer application for business and spouting the slogan "The network is the computer." You guys are coming from two completely different places, and two dramatically different generations.

JACK WELCH: So I'm basically the Neanderthal--

SCOTT MCNEALY: ...among other things...

WELCH: And this dynamite stud has had the thing going for years. Is that what this [interview] is all about--

No, it's more that your two worlds are converging, and your growing friendship is highly symbolic. Scott's on the GE board now, and he's obviously a new kind of character for that body. And in the meantime you have transformed into "e-Jack," spurring GE to become a leader among traditional old-economy companies in embracing the Internet. But we'll get to that later. Instead, let's start with an easy question for you, Jack: When did you learn how to type--

WELCH: In high school. A-S-D-F-space-J-K-L-semicolon-space. It's a terrible thing to say, but I never really used it much until my tenth wedding anniversary in Mexico two years ago. That's when I really started using the Internet. Since then, I've practiced typing on the weekends to get back up to speed.

So is this what a transformational CEO should do with his time--

WELCH: No, but it just makes things easier.

MCNEALY: I think it's a litmus test. If your boss can't type, doesn't do e-mail, and doesn't surf the Web, get out. If the boss isn't doing it, they aren't driving it down through the organization.

When did you first use e-mail--

WELCH: I'd say 24 months ago. My wife had a major impact on my game. She was all over this computer stuff. Having a second wife 17 years younger than you can get you in the game faster. I wouldn't advise that technique for everyone, but it worked for me.

MCNEALY: Must've been 1982, when we set up shop at Sun.

How many e-mails a day do you get--

MCNEALY: I get 200 or 300. I've got five direct reports right now, but I also have e-mail conversations going regularly with probably two-thirds of Sun's VPs, of which there are about 120.

WELCH: I get 40 to 50. I have about 20 to 25 direct reports, but I use e-mail to reach down into the organization, too. I just got an e-mail this afternoon from the fellow who is running our Spanish plastics factory. He's been having some startup difficulties in the past few months, and he was giving me the weekly progress report.

To emphasize a point, however, I still will mark something up on paper with a pen and then fax it. If I have a thing that's really in my blood, I don't find it as easy to communicate as passionately by e-mail. Somehow, the thickness of the pen makes me feel better; it makes it feel more meaningful to me.

MCNEALY: If you sent me a written note, I'd never get it. I'll go a whole quarter where I spend maybe four full days in the office. With e-mail, I'm never more than six hours away from receiving it, responding to it, and forwarding it on to people who can get something done. If you want action, you have to send me an e-mail.

Jack, GE has a "geek mentoring" program in which 1,000 Internet-savvy employees work closely with senior managers one-on-one to show them the ropes of using the Internet. What role did that play in helping you get comfortable with the Internet--

WELCH: Don't call them geeks. They are 1,000 young people who were relatively new to the company but who were very good on the Web. It was an idea I copied from one of our guys in Europe who kept telling me about his "mentor." This was the president of an insurance company, and I wondered, what did he need with a mentor-- Then he explained that he spent two or three hours with his mentor every week to learn the computer.

What we did with this was tip the organization upside-down so the senior people are all working with somebody junior. So we get all of the benefits and transparency of an upside-down organization. These guys all had mentors, they came in and did the stuff, and they learned a lot about new people, too.

Who was your mentor--

WELCH: She was a young woman in corporate headquarters named Pam Wickham. Pam now runs the GE Corporate Website. My secretary was also very helpful.

So what exactly do you do on the Web now--

WELCH: Besides e-mail, I look at financial services Websites. I go to Yahoo. I go to chat rooms and see what they're saying about GE. I'm tempted to jump in, but I don't. I go on almost every night to see what the gossip is. I go to the CNBC site, too.

MCNEALY: You know, the Internet is really three things: First, it is messaging--namely, e-mail. Second, it's a medium for transactions. Now everybody is all geeked up about business-to-business, auctioning, and other online transactions and trading. Third, it is becoming an entertainment medium. These are all quite distinct activities. You go to Amazon, E*Trade, or eBay to transact; you go to Disney's Go.com to be entertained. Me, I'm almost entirely messaging-oriented, so e-mail is what the Net is to me.

So, Jack, you clearly "get" the Internet as a consumer. More important, when and how did you have this epiphany that the Net really was important enough to prompt you to reorganize General Electric--

WELCH: I started realizing it in mid-1998. And after the Christmas season of 1998, when our office was buying everything online and not doing any shopping, and when my wife, Jane, was buying things for grandchildren on the Web, I realized that if I didn't watch it, I would retire as a Neanderthal. So I just started reading everything I could about it. I started jumping all over it.

One of the great things about GE is that when we turn, man, we know how to go. Take our industrial business, for example. We're taking all the databases of information we have on operating the turbines and engines and instruments that we make, and using them offensively. Now we can download improved software onto machines for people, or they can buy a service that lets them compare online the performance of their machines with similar ones around the world. A doctor, for example, can wake up in the morning and log on from home to see his previous day's CT scans, and compare how productive his CT-scan machine was vs. five other hospitals' in his area. We do the same with power plants. There are lots of things we can do with the Internet that gets us closer to customers. It fits our drive to provide better service perfectly. Growth in our service business is going through the roof, and I think that e-business is driving a lot of that.

There's a phenomenon I've started noticing that I call "the naked truth." When you use the Internet to open yourself up to more direct dealings with your customers, you've got to be pretty good at the nuts and bolts, or your flaws will really show.

MCNEALY: Yes, it gets a lot more Darwinian out there.

WELCH: But that's why it plays to a company that's got Six Sigma-quality programs, and rock-solid fulfillment capability, and reliability, and brand. Once we got to understand that, we realized this Internet thing is the easiest part, not rocket science. It's hardest for the guy who doesn't have the fundamentals down.

MCNEALY: Jack's right. Dot-comming your business is not all that hard. But it is scary, because it really does expose you to a real marketplace. That marketplace is a bid-ask thing, not a here's-my-price-list-and-long-term-contract thing.

WELCH: Comfort is not the issue anymore. The issue for us is that we don't need something between us and our customers. All people are trying to do with these dot-coms is to jump into the middle between the supplier and the customer. When you think about the relationships between supplier, distributor, and customer, many of these new businesses want to x out the traditional middle guy and be a new broker in the center. Our job is to be the ones that make that center guy evaporate. We don't want to pass our products through somebody else if we don't have to.

Early on, we did a couple of things: We used some Websites to distribute accessories--not really important stuff--that had always required messy, extra distribution steps. Before we knew it, that site was doing commerce with our customers on bigger-ticket stuff. An example would be medical equipment supplies--things that weren't critical but that customers needed to replenish from time to time. We could use that contact to talk about trading up to more sophisticated equipment. That helped us learn the big lesson: Never let anyone get between you and your customer.

Let's go back to bid-ask for a second. In this new, often auction-driven marketplace, how can you control the prices you set for your products--

WELCH: You better bring technology and added value to the customer.... Look at jet engines. Southwest Airlines makes a contract with us because we make the Boeing 737 engines. [Southwest CEO] Herb Kelleher knows where he wants his costs for the next 15 years. So we'd better have the technology and the models right to be there. We're online checking everything all the time, upgrading our systems, making their engines run more efficiently, helping them improve their processes online with us.

That's an information-based service as much as a product. And that gives you incredible leverage because that customer is going to find out exactly how he's doing all the time vis-a-vis his peer group. I think it's a better game. New MRIs [magnetic resonance imaging devices] cost 40% more than the machines they are replacing, but the images are better, and they are wide-open and less claustrophobic. Our new CT scanner, too, costs 40% more than the one it replaced.

MCNEALY: Time out! How come the price of everything he makes goes up and up, while computer prices keep dropping like a rock--

Seriously, the only problem with dot-comming your business is if you auction off your product without also auctioning off your purchase order. In other words, what you want to do is take all your purchase orders, auction them, and cut the price of supplies 10% to 15%. Then you go auction your actual product, with a better gross margin. But you have to do them simultaneously. What a lot of companies do is focus on how to get their customers a better price without also working the backward supply chain.

WELCH: But you desperately want to have enough technology in your product offering, so that you don't end up being in a commodity situation.

Where do you think we are as an economy as far as e-business goes--

WELCH: First inning.

And GE is out front-- Catching up-- Way behind--

WELCH: Against our competitive playing field, we're ahead of the game. Against an absolute standard, we're behind the game.

You're a GE director now, Scott. What's your appraisal of how GE is faring on the Net--

MCNEALY: I usually entitle my speeches "You're All Hopelessly Behind Dot-Comming Your Businesses." And after I get everybody depressed, I tell the old "60-foot-tall-Internet-bear-in-the-woods" story, which goes like this: There's this big Internet grizzly charging down the path at you. So you stop and put on your tennis shoes so you can run faster. That's sort of what Jack's doing at GE. A competitor might warn that there's no way GE can outrun that bear, and that may be true. But Jack's reply to him should be, "I don't have to outrun the bear, I just have to outrun you." And I would say that GE very clearly is outrunning the other traditional hikers in its businesses. The fastest elephant is a very good thing to be.

But at the same time, you have to worry about death by a thousand cuts, which is also what the Internet is all about. There's not going to be one big thunderbolt that kills you. If you don't dot-com your business, if you don't put your employees online, if you don't put your customers online, if you don't put your service data online, each one of those things will come back to get you. Most of these thousand cuts are self-inflicted.

As a GE board member, one of my jobs is to yell, "Fire!" Because the whole economy is on fire, in every way you can imagine. But it's not a big bonfire, it's lots of tiny Bic lighters everywhere.

Scott, what do you think of the big announcement by the Big Three in Detroit that they're going to try to set up an online parts-clearing house--

MCNEALY: I don't know the Detroit example well enough to know anything other than it's going to run on Sun equipment. I do think exchanges make a lot of sense. But from Sun Microsystems' perspective, I don't care. I'm an arms dealer, and I'll sell to you no matter what. I don't care whether you buy a Sun server and iPlanet software to run your auction on an exchange in your own company, or if you do a big one on the outside.

But customers do have to make a strategic choice--namely, Do they want to outsource purchasing and control over their own supply-and-demand curves, and thereby put another layer between themselves and their customers--

I'm not talking about routine things. For instance, at Sun, would we mind outsourcing maintenance, repair, and operations stuff like buying paper, pencils, tables, tissue paper-- I don't care. That's not strategic.

But we would never want to outsource buys of disk drives, DRAM, and processors. And where would I want people to go to buy used Sun equipment-- Sun.com, of course. Why would I want to lose touch with my original customer who is getting rid of it-- And why would I want to lose touch with my new customer who's buying the used equipment--

But there must be a reason that Detroit and the aerospace industry are creating purchasing exchanges for the key parts and components for their businesses.

WELCH: My guess is that some of us older companies think this Internet stuff is more difficult than it really is.

MCNEALY: That's why there are all these B2B startups. They're creating platforms and services for those that don't want to build it themselves.

WELCH: But I don't get it, Scott. Why would you hand somebody that responsibility-- It makes no sense to me.

MCNEALY: This is how the networking of business has changed everything. All of a sudden, demand curves are more important than supply curves. It's confusing for us, too. In the whole history of Sun we have never really known what the demand is, what the elasticities are, or what the "right" prices are for our equipment.

But when you go to a genuine bid-ask, Web-based system, inventory goes to zero and you've got a real demand curve. You know whether the ad campaign is working or not. You can actually test an ad campaign to see if the demand curve actually moved. That's going to challenge Madison Avenue, because all of a sudden we're going to find out that most ad campaigns are mainly for employee morale. Think about that....

If we print that comment we'll have to put a little editor's disclaimer beside it.... Scott, let me ask you the mirror question to one we asked Jack a while ago: When did you first recognize that the Internet really would be the killer app of business--

MCNEALY: The answer to your question was that Sun as a company was born to be an Internet company almost 20 years ago. Every computer we've ever shipped has been an IP [Internet protocol] computer. Every single one since 1982. As a company, we Web-ized everything first, and put everything in the browser and got totally networked, and now we're going back through and creating a learning organization. So I think you really ought to be asking me the opposite question: Namely, when did Sun get Jack's religion, and the leadership development that has made companies like GE so strong for so long--

Okay, when--

MCNEALY: I would say only in the past three years, about the same time Jack was catching on to the Net. Early on, we sold computers to engineers, who were the kinds of folks who were happy to buy a ten-speed bicycle and pay extra just to be able to put it together themselves. We could sell them just about anything.

Then we went into the enterprises, selling to the mission-critical and life-critical parts of the enterprise, and it got harder. Quality doesn't matter so much as constant availability. You can ship junk, but once it gets up and running, it's got to stay running. That's our biggest challenge. Quality is important, but when you provide service it's more an issue of availability and reliability.

This has been a hard thing to get across. When I asked people to respond to this service-oriented mentality, the levers of the organization seemed rubbery, and the bells I was trying to ring sounded like they had socks over them. That's when I decided to study the big-elephant companies that can really dance, to see if I could learn a new trick or two. Being CEO of Sun today is so different from being CEO of a much smaller company.

Is that because you could see Sun broadening into something more diversified and more like GE--

MCNEALY: No, it was the sheer size of the organization. The important measure of size in a company isn't revenue or lines of business, it's the number of people.

Let's bring this conversation full circle now. Jack, how and why did you get to know Scott--

WELCH: A few years ago, Golf Digest published a list of CEO golfers, and Scott was ranked No. 1 and I was No. 2. Scott sent me a challenge, saying "If I'm going to be No. 1, I want to be sure I'm No. 1." Isn't that right--

MCNEALY: What I did was write him, in a thinly disguised ruse....

To sell him servers--

MCNEALY: ... no, no, no, to meet my business hero. I've been a Jack groupie for years. I knew he played golf, and I figured this was a way to get some real face time. Not to mention that GE resells a couple of billion dollars' worth of Sun equipment and is a huge end-user customer. We're also partners--our products go into GE medical instruments, and a lot of GE's engineers and designers use our equipment. So this was a good guy for me to get to know.

So I challenged him. It was so easy for him, it was like chumming for sharks in a bathtub, because it was all on his terms. I think the actual challenge was "Jack, you name the place--any time, anywhere, mano a mano, and we'll settle it once and for all." I got a phone call in about three nanoseconds, and we set the date.

And then you took a dive--

MCNEALY: Uh, no.

WELCH: You immediately got to the cynical question that I have never gotten to over these two years. I was more innocent. I actually thought he wanted a golf match. I did win, though. [He points to a trophy commemorating his victory.]

We're journalists. We have to ask cynical questions.

MCNEALY: Now that I think about it, the very first time we met was ten years ago, just when the Big Bertha came out. I remember because I was trying to sell computers to someone at GE, and Jack walks in with a shrink-wrapped Big Bertha driver and says, "You gotta have one of these! Here!" and hands it to me. So I ended up walking around the rest of the sales call holding a driver.

We can understand why you are good golf buddies, but Jack, what did you want when you invited Scott to join the board--

WELCH: Scott brings GE a whole irreverence. He's an enormous pain in the ass sometimes, yelling and screaming. He makes our guys know it's okay to be informal. We pride ourselves in being direct and informal, but he's much more informal than we are. He talks about speed in ways we don't usually think about it. He brings urgency, a view of where things are going with IT.

I wanted all the brains and enthusiasm of another generation and an industry that's changing the world. I think getting Scott was a big deal for our company. And having Scott as a role model for the young managers in our company is fantastic. So for us, it's a real coup. He's young enough to relate to our young people in a totally simpatico way.

So, Scott, do they actually try to get substantive input from you about their infotech strategy--

MCNEALY: They don't get a chance to ask. I have a hard time not offering suggestions. There's a seniority thing at GE where the rookie sits at the end of the board table. I think I'm probably the noisiest rookie they've seen in a while.

WELCH: Kenneth Langone [chairman and CEO of Invemed Associates] is right up there with you.

Are your companies, in effect, exchanging viruses--

MCNEALY: Actually, I think we've gotten way more out of Jack than they've gotten out of us. We're becoming a more competitive company by learning how to take advantage of after-the-sale revenue opportunities through services; we're learning about the boundaryless way to operate inside a company; we're driving through Sun this whole Six Sigma approach to quality.

Hold on a second. Jack, for our readers' sake, what exactly are we talking about with Six Sigma, and how did you get started with it--

WELCH: Larry Bossidy [CEO of AlliedSignal] and I equally hate quality programs. We thought they were just slogans. Larry called once, though, and said, "I've found the answer to getting costs out, and it's this Six Sigma program." So I had him come and talk to our guys. He made an impassioned speech. I was in the hospital with a coronary bypass. When I came back, everyone was really excited about what Larry had said. This was September of 1995. Then a guy came in with a computer and all this statistical analysis, and it was like a dental procedure. But we went with it. Six Sigma allows you to go to your customer and find the critical quality issues. We never introduce a product today that hasn't been thought through from the customers' eyes.

Take the open MRI device that we introduced. Why did we do that-- Because people hate the claustrophobic nature of the experience of having an MRI scan. And the doctor wantsthe same image quality as with the old closed devices. We took care of both. Boom. It's taking off. Hospitals are buying them like grapes.

Six Sigma allows you to think about projects by keeping in mind what's critical to the customer in terms of quality.

MCNEALY: What he did was create a common language for understanding what customer requirements are and how you go about meeting them. What I'm finding in a company of 37,000 employees is that there are different dialects being created across the company, which makes it hard to get common reporting. You can't compare one organization with another. We have a saying inside Sun: "If you can't measure it, you can't fix it." You certainly can't measure it if you can't come up with the same labels for the axes of the charts and graphs. One of the advantages of getting a program going like this is that you get a common language.

WELCH: That turned out to be true for us. When we go to Thailand and review a business, they're talking the same language because they're talking Six Sigma.

MCNEALY: Then you get your cross-pollination and share knowledge across the organization. You can move managers around, develop them, broaden their experiences. One reason Jack has done so well is that he's able to move his people around.

WELCH: It's not just a common language, it's a common value set.

So, Scott, what else are you learning from Jack--

MCNEALY: Jack has seen a movie I haven't seen. All of a sudden we're at 37,000 employees, we're growing at 20%-plus growth rates ...the kinds of things that Jack has done, I'm beginning to have to deal with. It's a vast organization. You can't just call everyone into the lunchroom and stand on a chair and tell them, "Here's plan B." That's what I used to do.

It's a very, very different process. Jack has developed a learning organization that can spin on a dime, because he's got these black-belt, Green Beret-type folks infiltrated throughout the organization. So when the word comes down that this is the new initiative, away they go.

The other thing that's fantastic that Jack has done that I'm trying to do at Sun has to do with this: The bigger the boat gets, the more crisp, clear, and sparing you need to be about picking strategies and ideas to pursue. For GE, globalization was one, building a boundaryless organization was another, product service was another, Six Sigma quality, and now the Web. There have been just five companywide initiatives in Jack's whole career. My folks will tell you that I've got five initiatives per meeting. So one thing I'm learning to do is to step back from spewing an idea a minute to focus on driving higher-level issues. For us, that's things like chip development, or availability as opposed to quality. I'm going to pick very few fights going forward, and I'm going to win them. That's the best thing I've learned from Jack.

WELCH: The bottom line to me is that Scott's just a good guy. I like to play golf with him; I like to pal around with him. I have great respect for him, and I think he's full of baloney sometimes, but that's a pretty good deal. I like the hell out of him.

MCNEALY: That's another important message I've learned from Jack. We all work so hard, and if you aren't working with people you enjoy, you've got a bad life. People ask me if I'd ever go to another company or a startup, and I have to say, I love the people I work with, so I'd find it very hard to walk away. They're the greatest, most fun, interesting people I could hope to be around. Any CEO who doesn't say that--well, he's the one who can fix it.

One last question: Has having Scott around changed your thinking at all about succession at GE when you retire next year--

WELCH: I have nothing to say about that.

Won't you even say "obviously"--

WELCH: [silence]

MCNEALY: Jack has me very well trained. I've got nothing to say about that either.

Back to top

Solar Power

Forbes
January 22, 2001
Daniel Lyons

Death lurks just around the corner for Sun Microsystems. That's what rivals have insisted for nearly a decade, certain their low-cost machines running Intel chips and Microsoft software eventually would overwhelm Sun's pricey Unix servers.

How annoyed those naysayers must be. Sun raced past them by taking center stage in the Internet revolution. Its revenues could eclipse $20 billion this year, doubling in just three years and tripling in the past five; since 1996 net income has soared fivefold to a projected $2.5 billion or so. In the September quarter revenues soared 60%--three times as fast as two years ago, when Sun was only half as large. Faster, in fact, than at any other time in the past decade. No wonder Sun's stock price has increased tenfold in the past five years.

But at long last, Sun's doubters have new reason to feel hope. High tech is in a slump. As the dot-com crash ripples through the digital economy, the company that "put the dot in dot-com" is vulnerable. In every direction, destruction abounds: PC companies are falling short on sales; Internet-access firms and telecom upstarts are struggling; even the once-mighty Microsoft has been reduced to single-digit growth. Some of the Web sites that contributed so generously to Sun's growth are going kaput. Web-design and hosting firms are getting slaughtered, their stocks down 90%or more. Old Economy stalwarts are tightening their belts and may spend less on the Web, given the now-idle threat from dot-com has-beens.

Investors reckon Sun can't escape the devastation, and they have pounded the company's shares accordingly. From September through December the stock plunged more than 50%, to as low as $26. Even after that jolt, Sun's shares still trade at a pricey 50 times trailing-12-month earnings. If trouble erupts, Sun's stock could fall even more; if investors deemed Sun worth the same multiple as, say, a Dell, the shares could fall by more than half yet again.

Even Sun's relentlessly upbeat chief executive, Scott McNealy, is bracing for pain. In a Dec. 19 memo to employees, McNealy warns of a slowing economy and says Sun will pull back on spending and hiring plans.

But that amounts to mere tweaking on a massive growth engine. McNealy and Sun's president, Edward J. Zander, insist they see no short-term threat and say that in the long run Sun could be set to boom like never before. The company has always thrived on moxie and bellicose tactics, by being smarter, faster and meaner than its rivals. It aims to reinvent itself yet again; it went from a maker of workstations for engineers in the Eighties to a producer of Unix servers for business in the Nineties to a purveyor of Web engines and software for the next decade.

"We have this uncanny ability to seize disruptive technologies and just go maniacal on execution," Zander says.

He joined Sun in 1987 after a career at two rivals that Sun later pounded into dust--Data General (now a unit of EMC Corp.) and Apollo Computer (which disappeared long ago into Hewlett-Packard). Zander, 53, has been McNealy's number two since 1998, and he is as low-key, calculating and restrained as McNealy is flashy, transparent and scrappy. McNealy,46, sets the big vision and picks fights with powerhouses like Microsoft. It is up to Zander, who has every Sun division reporting to him, to actually make it all work.

To hear the two of them tell it, the Internet build-out has hardly begun. Others share that opinion. Annual U.S. spending on Net infrastructure, at $200 billion in 2000, could grow 75% by 2003, says Jupiter Research. "This is the biggest equipment opportunity in the history of anything," McNealy says with characteristic hyperbole. "If we were writing a book, we would not even be at the prologue. We're barely at the copyright page."

More like the first big plot twist, but let's go with it. Even in the short term, Zander and McNealy claim to have no fear of a slowdown. They boldly (or unwisely) say Sun might even gain in a downturn, as companies shift spending away from nonessentials like PC upgrades to spend more (supposedly) on Internet projects that can stoke sales and cut costs.

"I think there's going to be a shift in capital spending, and it's going to be lucky for us," Zander says. "Idon't care who you are--if you're General Motors or Deutsche Telekom, you're going to keep investing in Internet infrastructure and dot-comming your company, because that's going to let you gain global advantage and be more competitive."

Zander, Brooklyn-born and bred, says his days at Data General and Apollo taught him how quickly a tech company can stumble. He studies organization charts and shakes things up regularly. "I've done a reorg every year that I've been here.Idon't reorg defensively, I reorg offensively. Just as we get comfortable, I'll change things."

Five years ago Sun was among the first to spot the huge market ahead as businesses began linking up with the furiously proliferating Internet. Sun optimized its computers for serving up Web pages. It also created Java, a programming language it distributed free of charge, establishing it as the lingua franca for writing Web applications. Those things, plus a marketing campaign based on the slogan Zander penned on a cross-country plane trip--"We put the dot in dot-com"--propelled Sun into the top spot among Internet arms dealers.

Visit giants like America Online and Ebay, or Web-hosting firms such as Exodus Communications and Digex, and the same linear image shows up:row after row of sleek, rack-mounted computers punctuated by Sun's diamond logo.

That owes to a daring bet Sun made five years ago. Hewlett-Packard, IBM and others, succumbing to what seemed like the inevitable rise of Microsoft's Windows and Intel chips in high-power servers, began turning away from the Unix operating system to develop "Wintel" systems. Sun resisted and focused its research spending on its own brand of Unix, Solaris, and its own in-house chip, the Sparc.

"The analysts were committing us to the funny farm," Zander says. But Sun got lucky: Microsoft's operating system didn't evolve fast enough and robustly enough to handle the heavy lifting required of a Web server. Unix servers became the backbone of the Web, with Sun making most of them.

"You know those tuna boats where they haul in these huge nets, and the guys stand there with hooks, hauling them in as fast as they can--That's what it's like here these days," says John McFarlane, a Sun executive vice president.

In the $32-billion-a-year market for Unix servers, Sun sells more than HP, IBM and Compaq combined.It ships 48% of units worldwide and has a 39% share of revenue; the number two player, HP, has a mere 16% of units and a 23% share of dollars. Prices run from $10,000 to more than $1 million, with the most powerful able to handle several thousand transactions per second.

At Sun, McFarlane runs a division devoted solely to selling to Web hosters and other service providers. Industrywide, the provider business could more than quadruple to $19 billion in annual revenue by 2003, says research firm IDC. McFarlane's shop, formed only in 1999, already has surpassed $1 billion a year in sales.

Sun has cemented its lead among service providers by doting on them unendingly. Product managers in the new division spend at least one week per quarter at a service provider's site. Sun even steers potential customers to service providers by paying its sales reps higher commissions when customers agree to let a service provider run their Sun systems.

"Sun really understands how important our business is to them,"says Robert Patrick, a vice president at Web host Digex, which posted 190% revenue growth in the September quarter and sees no pullback. "Our big customers are still rolling out business-to-business sites and new supply chain projects. They're all looking for ways to do things more cost effectively.We're booming,"Patrick says.

Likewise, Enron's broadband division has committed to spend $350 million on Sun servers over the next five years and isn't scaling back, says James Crowder, an Enron vice president. Sprint Corp., also a Sun customer, plans to spend almost $2 billion in the new year to build 11 new data centers for hosting corporate Web sites. It hopes to erect 7 more centers in 2002. "We're seeing tons of demand,"says Keith Paglusch, president of Sprint's E-Solutions division.

But Sun could stumble if service providers scale back in a marketwide slowdown. This is where high-tech's genetic predilection for optimism kicks in:Sun and its Web-host customers argue they could actually benefit in hard times. Corporate clients could decide they can save money by farming out their Web tasks.

"When times get bad, people turn to outsourcing to save costs," says Ellen Hancock, chief executive at Exodus in Santa Clara, Calif. Then again, if the outsourcers prevail, it could one day undercut demand for new gear because fewer businesses would be buying their own systems.

True, Exodus' business is roaring. It runs Web operations for 3,300 companies, including General Electric, for which it handles ten divisions and soon will add more. Exodus' sales surpassed $240 million in 1999 and ballooned past $800 million last year. This year sales could more than double, Hancock says. But even she has delayed some projects because of the shaky market ahead. "We're getting a little more cautious,"she says.

More alarming for Sun, cutting back can mean switching from Sun to cheaper Wintel systems. Sun used to supply 90% of the infrastructure at Exodus; now 50% of the Web host's gear comes from the Windows-and-Intel camp. Other customers similarly are using small, cheap Intel servers to handle less demanding tasks, such as letting users view static Web pages at the front end where visitors first enter.

Sun had assumed companies would prefer to run a few big Unix servers rather than rely on hundreds of smaller Wintel systems. Wrong: On the low end (systems costing $100,000 and less), cheap systems based on Microsoft's Windows NT now outsell all Unix machines for the first time, in both dollars and units. Worse yet, the low end is already bigger than the high- and mid-price markets combined, and it grows faster. By 2004 Windows NT servers will outsell low-end Unix servers 2-to-1, says research firm IDC. "The tide is rising against Sun," says Michael Lambert, a senior vice president at Dell. "Every year we keep pushing up against them from below."

At long last, Sun is striking back. It is taking aim at the Intel-based bottom feeders like Compaq and Dell and trying to attack them from below. Last month Sun closed the $1.3 billion acquisition of Cobalt Networks in Mountain View, Calif. Cobalt makes "server appliances," which are even cheaper than Intel-based servers and run on Intel-clone chips and Linux software (a free version of Unix).

The Cobalt deal is the latest in a string of $2 billion in deals in the past year at Sun, which had always grown almost entirely internally. "We've spent more on acquisitions in the past six months than in the previous 18 years," says Jonathan Schwartz, the Sun vice president who handles acquisitions.

Sun also faces formidable competition at the high end (machines priced at $1 million and more). IBM and HP discount heavily and sell new Unix superservers that outperform Sun's biggest entries, while costing less. Sun's top-of-the-line E-10000 server, which shattered world records for performance in March 1999, now ranks tenth, says the Transaction Processing Performance Council, an independent group. Sun disputes that claim. Still, Sun has had to discount to stay in the game, contributing to a four-point slide in its gross profit margin in the most recent quarter. Another snag: Delays in rolling out its next line of big servers.

Some skeptics still doubt that Sun servers are sufficiently brawny and bulletproof to power a worldwide enterprise. Sun's Zander recently hosted a daylong session for executives of Charles Schwab & Co., hoping to grab a piece of its business. But Schwab uses little Sun equipment and confines it to nonessential applications. The important stuff will stay on mainframes and big Unix servers from IBM.

"I'm not going to put high-volume, mission-critical trading applications on Sun boxes until they prove they can handle it," says David Dibble, a senior vice president at Schwab.

Another obstacle: the reliability bugbear. Last year Sun wrangled with a nasty glitch that caused its high-end servers to crash without reason. Some customers were kept waiting weeks, or even longer, for repairs.

Still, Sun is making strides in high-end corporate systems, a market IBM has dominated for decades, says Scott Thompson, chief technology officer at Visa USA. A few years ago Visa USA took a cautious first step with Sun computers, using them only for nonessential tasks. Today Visa runs its core system on Sun machines, which process 35 billion transactions a year. "Sun is really making an effort to be world-class in the top end of the enterprise market,"Thompson says.

Sun has improved the old-fashioned way--by stealing pages from IBM's playbook. In Decemberit introduced software aimed at making high-end systems more crashproof, copying ideas long used in IBMmainframes. Like IBM, Sun also places its engineers on-site in customer data centers. Merrill Lynch, which runs its trading floor and e-commerce on Sun servers, has a team of Sun engineers in-house. At FedEx in Memphis, a Sun team helped develop the network that tracks 45,000 FedEx trucks and planes.

Zander and McNealy believe, however, that ultimately most companies won't need such on-site doting because they will hand off their tech operations to where Sun is king. They envision customers purchasing computer power from service providers just as they get water and electricity from utilities.

Sun's goal is to become a one-stop supplier for those utilities. Core to this vision is a bundle of hardware and software--what McNealy calls "a big freaking Web-tone switch"--that he hopes to deliver the way Nortel delivers switches to telcos. McNealy says customers are sick of having to cobble together Web sites using products from dozens of suppliers. "They want one throat to choke."

McNealy's dream machine would require no systems integration, no figuring out what software goes with what, no writing device drivers and jiggering and configuring--just drop it on the floor, flip the switch, there's your Web tone.

Easy to say; not so easy to do. Sun is far from having all the pieces. Its Web software offering, iPlanet, is a hodgepodge, some of it developed by Netscape (a division of America Online, which has a joint venture with Sun)and some by Sun. The software has only 9% of the market versus 24% for IBM and 24% for BEASystems.In storage, Sun holds a mere 6% market share, ranking fifth among all vendors.Most customers who buy Sun servers still look elsewhere for their storage. In services, Sun fields 3,500 people, puny compared with IBM's 50,000 consultants.

Sun can't make any of its vision come true if the service providers it so dotes on suddenly halt their spending and corporate clients stop expanding their online efforts. How bad could it get if Zander and McNealy are wrong, and Sun really does take a hit this year, as Wall Street already expects--

The good news is, Sun is a tight ship. Yes, the company hired 10,000 new employees last year, bringing head count to 42,500. And Zander expects to bring another 10,000 on board this year. But sales, general and administrative costs represent a smaller percentage of revenue today than they did three years ago (25% now versus 29% then). Operating margins (profit after costs but before interest income) were 14% in September, up from 9% in 1997.

Zander is also placing smart long-range bets that will take Sun into new markets. Add a dash of special software, for example, and a Sun server can act like a phone switch. So-called soft switches, costing about half as much as traditional circuit switches, are the next big thing in telecommunications. And Sun is the server of choice for companies that are building them, like Nortel and Xybridge.

Sun may be able to weather a tough year and emerge stronger for the next phase of theInternet build-out. The Internet today has 340 million users, almost all of them connecting over PCs. Someday there will be billions of people and gadgets on the Internet--cell phones and wireless devices and game consoles and automobiles, even household appliances--all fueling demand for more Sun servers.

"New burdens,"McNealy says. "Everything with a digital or electrical heartbeat will be hooked to the Internet. Look at all the equipment we're selling today, and we haven't even started."

Sun's stock sells at about 50 times trailing earnings--less than half what investors pay to own shares of EMC or Cisco, the other top suppliers of Web infrastructure gear. No doubt McNealy considers that the bargain of a lifetime. If Sun can ride out the coming storm he may just be right.

Back to top

McNealy's cold feet and other tales of Sun; Co-founders recount Sun's early history, including CEO's reluctance to join and attempted mergers with Apple.

January 12, 2006
Stephen Shankland
CNET News.com

A correction was made to this story. Read below for details. MOUNTAIN VIEW, Calif.--Sun Microsystems CEO Scott McNealy had to be wined and dined at a Silicon Valley McDonald's before he gave up his reluctance to help launch the workstation maker in 1982, according to one of many tales the company co-founders recounted on Wednesday.

McNealy joined Sun's other co-founders, Vinod Khosla, Andy Bechtolsheim and Bill Joy, at a panel discussion at the Computer History Museum here to reminisce about the server specialist's past and prognosticate about the future.

Khosla said the McDonald's meal took place just after he and McNealy met with venture capitalists and got Sun's first funding commitment. "We went out and sat in the parking lot. Scott said to me, 'I don't know if I really want to do it.' So I took him to an upscale dinner at McDonald's on Page Mill Road" in Palo Alto, Calif., he said, where he put the screws on McNealy to resign from his $40,000-a-year job at Onyx Systems.

"Vinod asked me, 'When are you quitting--'" McNealy recounted. When McNealy balked, Khosla countered, saying: "'You can't back out on me now. You're a founder.' "I said, 'Oh, OK.' It was that quick," McNealy said.

Khosla left Sun in 1986 to become a general partner at venture capitalist firm Kleiner Perkins Caufield & Byers, and Joy followed suit in 2005. Bechtolsheim left Sun in 1995 to found gigabit Ethernet start-up Granite Systems, later acquired by Cisco Systems. But he rejoined the company in 2004, when Sun bought his next start-up, Kealia, to provide the foundation of its new "Galaxy" line of x86 servers.

The Apple connection

It's no secret that Sun once tried to acquire Apple Computer, but Joy noted that it wasn't the only near-union between the Silicon Valley companies. "We almost merged with Apple two other times," he said.

There were other alliances with Apple that fell through, Joy added: an attempt with Microsoft and the Mac maker to create a common file protocol; an attempt with Apple to create a merged user interface; and an attempt to persuade Apple, when it was moving away from Motorola's 680x0 processor family, to switch to Sun's Sparc processors rather than the PowerPC chips it ultimately chose (and began abandoning this week with Intel-based Macs).

"We got very close to having Apple use Sparc. That almost happened," Joy said.

In total, "there were six very, very close encounters" with Apple, he noted. That none of them worked out was a "personal disappointment" said Joy, who spent years as Sun's chief technology officer.

"The tip toward the public space being much less private is one that's hard to fight." --Sun co-founder Bill Joy

McNealy added that he went to Steve Jobs' house to try to hammer out the user interface agreement. The Apple co-founder and CEO was "sitting under a tree, reading 'How to Make a Nuclear Bomb,'" with bare feet and wearing jeans with holes torn in the knees, McNealy said. The interface work, though, "never went anywhere," he said.

Khosla also lavished praise on Jobs, who he said was a role model, along with Oracle CEO Larry Ellison and Intel's former CEO Andy Grove. Jobs is the kind of person "who passionately, religiously believes his own ideas. No matter what anybody else says, he's going to push them through," and that determination and self-confidence is in large part why he succeeds in doing so, Khosla said.

McNealy has praised Jobs on occasion, but he acknowledged on Wednesday that he doesn't have time to listen to his own iPod and forecasted doom for the popular digital music player. The right place to store music is on the network, where it can be accessed by many devices, he said, much like the right place to store voice mail is on a central server.

"Your iPod is like your home answering machine. It's a temporary thing," McNealy said. "It's going to be hard to sell a lot of iPods five years from now, when every cell phone is going to be able to automatically access your library wherever you are."

Correction: This story misreported the amount of money Sun takes in every year. The company's annual revenue is approximately $11 billion.

Of course, Sun is the company that wants to sell the servers that could store such material, the software that could govern access rights and the Java software to power cell phones. Its various product lines are reflected in the company's tagline: "The network is the computer."

The tagline was invented by Sun's chief researcher, John Gage, whose hiring at Sun was a condition that Joy placed on his own employment. Gage invented the phrase while headed to Japan for a trip, he said, but the second half of the expression never caught on--"The machine is the manual," meaning that using a computer should be a self-explanatory experience.

"We are the last server company that's survived." --Sun CEO Scott McNealy

Another close call with Apple came in 1991, when for an April Fool's Day joke, vice presidents at Apple and Sun went to the other company's CEO staff meeting, Gage said. The executives wore masks made of 8-by-10-inch glossy photos of their counterparts from the other company.

Apple CEO John Sculley was not amused when Bechtolsheim arrived wearing a mask. "Sculley looked stone-faced. Maybe he thought some terrorists were trying to take over," he said.

But the founders didn't restrict themselves to the past. McNealy reiterated his infamous quotation from the 1990s--"You have no privacy. Get over it"--and forecast that today's trend of cell phone cameras will transform into tomorrow's trend of using the devices to record everything that happens.

"They're going to put this (mobile-phone recording device) right here (in a shirt pocket) and walk around all day and record it," McNealy said.

Joy had a similar view. "Technology and privacy are on a collision course," he said, because ever-cheaper processor technology means it's constantly getting easier to create sensors and collect data. "The tip toward the public space being much less private is one that's hard to fight. It has Moore's Law on its side."

Early days at Sun

As it turned out, McNealy wasn't the only founder who didn't leap at the chance to join Sun.

Khosla had to convince Bechtolsheim that he was more interested in the computer engineer than the software he was licensing from him. However, Bechtolsheim was reluctant because his licensing business brought in $500,000 a year and Bechtolsheim had a year to go before finishing his doctorate, Khosla and Bechtolsheim said.

Convincing Joy wasn't easy either. McNealy said he and others piled into a rattletrap of a car to visit the Unix wizard across the San Francisco Bay in Berkeley, Calif., but they got the cold shoulder. Later, he asked Joy, "Why didn't you talk to us--" and was told, "I was waiting for top management to show up," McNealy said.

Fellow computing expert Bechtolsheim and Joy hit it off, though. Joy had a collection of six VAX 750 computers from Digital Equipment Corp. "It looked like a computer center, but they were actually my machines," Joy said. "I took him in the machine room. I walked up to a minicomputer, turned it off, pulled one of the boards out and said, 'Here, look at this. It's a very early system from DEC.' That was our way of bonding."

DEC was one of Sun's prime targets, and the company's success came at DEC's expense. "The VAX 750 had an 80MB disk drive, a couple megabytes of memory" and was able to process about a million instructions per second, Joy said. "Andy's design from standard components, once we had the (Motorola) 68010, was the same thing but at one-tenth the price. That was really a revolution in capability." Joy, who had earned a reputation as the power behind Berkeley Software Distribution, which modernized AT&T's Unix, was a major draw for early customers. McNealy handled purchasing calls in the company's early years using a phone with four or five lines. One early conversation went like this:

Customer: "Is Bill (Joy) onboard--"

McNealy: "Yeah."

Customer: "I want two of whatever you got. What are you selling--"

Among McNealy's other tasks was writing the company's first accounting software using the Unix vi editor--though he initially used Onyx systems and not Sun's own. He also oversaw "burn-in" tests to weed out flawed computers, turning off the office air conditioning to find any machines that couldn't endure the overheating.

Gage was the company's original salesman, McNealy said, and kept track of prospects on pink slips of paper. When calling back customers, "He'd organize pink slips by time zone. He'd start at the left side of his table and move right."

Joy marveled at Gage's system. "Every once in awhile, John used to take all the slips and throw them in the trash. He said, 'If it's important, they'll call back.'"

The early years left an impression on McNealy, who still can recall how many millions of dollars in revenue the company garnered in each early year. "It was 8.5, 39, 110, 210, 450, then a billion. Then I forget after that," he said.

Sun has annual revenue of roughly $11 billion now but has struggled financially in recent years. McNealy is confident that Sun will survive, though, because innovation matters and it's not easy for competitors to get started.

"Barriers to entry in our business are big, because it takes a lot of capital to do what we do...(Hewlett-Packard) kind of checked out, in that they don't do microprocessors, operating systems, the software stack. At some point, you no longer are a car company, you are a car dealer," he said. "The major research and development is being done by (Advanced Micro Devices) and Intel, Microsoft, Sun and IBM," he said.

And McNealy is convinced Sun has staying power. "There haven't been any other companies that started after Sun," he said. "We are the last server company that's survived."

Back to top

 
  
 

ON THE RECORD

Read Sun's News Blog
 
Press Contacts
Frances Freyberg
Sun Microsystems, Inc.
+1(650)352-4770
frances.freyberg@sun.com

Carolyn Rohrer
Sun Microsystems, Inc.
+1(415)294-5084
carolyn.rohrer@sun.com
 
Executive Bios
»   View Executive Bios
 
Company Info
»   Corporate Facts at a Glance
»   Company Logos
»   Company Information