Ray Ozzie, Microsoft's new chief software architect, is on a mission. *
Photo: Lionel Deluy * The keynote speaker at this past summer's TechReady conference—a gathering of 6,000 or so Microsoft engineers from around the world—was the company's chief software architect, Ray Ozzie. This was not a routine appearance. Ozzie arrived at Microsoft in 2005, and the following year he inherited the title of CSA directly from Bill Gates. He was now the microprocessor of the Microsoft machine. But he had never addressed the semiannual conclave. His explanation? He wanted to wait until he had something big to show the troops.
But there's something else: Ozzie hates speaking in public. His idea of paradise is pitching his vision around a table near a whiteboard, where he can proceed conversationally and draw on his marketplace savvy, quiet confidence, and ability to scrawl out XML code on the fly. Auditoriums are something else. "I have high anxiety—massive, huge, tremendous anxiety," he says. "It's not a natural act for me." The infrequency of his public appearances has triggered murmurs that the guy in Gates' chair is afraid to face his public, like some sort of software Greta Garbo. "Where's Ray?" Microsoft observers have been asking, as Google grabs more headlines and Apple relentlessly mocks the company's shortcomings. Two-plus years into the job, there is still a bit of mystery to Ray Ozzie.
It is about time that one of the most significant figures in the personal computer age, the writer of Symphony and creator of Lotus Notes, emerges from the shadows. Time to reveal what he has been working on. And, most important, time to explain how the world's mightiest software company is going to remain relevant.
Not an easy job. Yes, Microsoft still rakes in the dollars from Windows and Office. But the stock has been flat for years. Microsoft used to be regarded with fear and respect—Lord Voldemort with market share. Now people downgrade their computers to avoid Vista, tech luminaries write blog posts with titles like "Microsoft Is Dead," and the public face of the company is the hapless loser in the Apple ads. Oh, and this year, after a 25-month transition, Gates, the once-omnipresent cofounder, left the building.
To get a sense of the gloom, check out the speakers who precede Ozzie at TechReady. First comes an executive who presents a preview of Microsoft's upcoming ad campaign—a protests-too-much response to those PC-versus-Mac ads from "that fruit company down the road." Then he describes the remedial work Microsoft is doing to get computermakers to tweak their products so a Vista bootup takes less time than a round of golf. The next guy reveals the not-so-encouraging information that he has taken a hard line on his children's pleas for iPods. They'll go without, he promises, as long as "my rent is paid by Microsoft." He doesn't mention where his kids go when they want to search the Internet. The odds, of course, are that they use Google, the rival that's even deeper inside the heads of Microsoft executives than Steve Jobs is.
The tone changes when Ozzie appears onstage.
If his insides are churning, his delivery doesn't show it. His demeanor is smooth, direct, gimmick-free. At 53, Ozzie is trim and fashionable—eschewing the Redmond standard-issue chinos and polo shirts for a dark suit, gray shirt, and no tie. His hair is burnished silver, as it has been for years. (Take away the stage fright and he could be the lead detective in a Law & Order spinoff.) Before Ozzie gets to the software demos that he hopes will revivify Microsoft, he addresses the lingering "Who is this guy?" question.
"You and I don't have much time together at events like these," he says, "and yet, given how much we need each other, it's probably good for you to get a sense of how I'll lead in my role as CSA. Unlike with Bill, you haven't had 30 years to get to know what makes me tick." To remedy that, he offers some quick bites about his passions—in a style suggested, he says, by the 140-character format in Twitter. (The fact that he references Twitter and not a Microsoft product is a statement in itself.)
I love software, because if you can imagine something, you can build it.
I love Windows, because without it there would be no PC. There would be no PC developers. There might not even be a Web.
I love the ubiquitous Web because of the connections that it opened up.
I love competition. But when we're behind a competitor, I hate it when we find ourselves just chasing their taillights.
Ozzie works in one of his two Redmond offices. A Microsoft outsider, the new chief software architect was handpicked by Bill Gates to retool the Windows giant for the 21at century. Gates calls Ray Ozzie "one of the top five programmers in the universe."
Photo: Lionel Deluy After the introductory Tweets, he gets down to business. At Microsoft, he says, there must be a shift from the traditional model of software to what he calls software plus services. As slogans go, it's not particularly catchy. But the sentiment is clear: Just packaging software, collecting the money, and then producing a new version a few years later (whether people want one or not) is no longer a sustainable business plan.
The relationship with customers must be constant and continuous. Instead of discrete onetime transactions, the money—whether from subscription fees or advertising—will flow constantly. For the user, everything will happen when it's needed, as if pulled down from a cloud. The metaphor has been around for years, along with the more recent spinoff, cloud computing. But the phenomenon is anything but ethereal. Billions of dollars are at stake.
According to Microsoft, one example of a successful service is Windows Update, which automatically installs patches and bug fixes on users' operating systems. Hotmail, like all Web-based mail applications, is also a service. Virtual Earth? A service. Software, but not from a box. Still, Ozzie draws the line at the idea that you can do anything and everything in the cloud, that every application can become Web-based, that the desktop is dead. Some things, he says, still require local computation, offline persistence, and the control that only one's own desktop processor offers.
This defense of the desktop dovetails nicely with Microsoft's historic strengths. So, while Ozzie actively evangelizes for the disruptive move to services, he's also saying that for many purposes the ideal software model is a hybrid: a heavy-duty application (known as client software) combined with an ongoing Internet service. A great example is Apple's iTunes, which you install on your computer and use as an offline media organizer but which also serves as an Internet app that lets you buy songs, stream music, and get recommendations.
In Ozzie's view, Microsoft must make this model the centerpiece of all its future efforts. The company must transform itself from a manufacturer that dumps out a big product every couple of years to a customer-obsessed enterprise devoted to continually producing, updating, and supporting a full panoply of services. In his speech, Ozzie puts it this way: "When packaged software ships, services go live. What was our end is now the beginning. The gold disk"—from which all retail copies of a new piece of software are made—"is now the grand opening."
Ray Ozzie was a University of Illinois sophomore in 1974. At that point, Ozzie unveils the new products that he's been laboring over for more than two years: a top-secret set of initiatives designed to make Microsoft as dominant in the cloud era as it was in the days of the desktop. First up is a new operating system for Web-based applications, codenamed Red Dog—it's Windows for the cloud. (See Editor's update below.) Then comes a demonstration of Live Mesh, which will allow people to seamlessly synchronize all their information with as many people and places as they want, across as many devices (computer, phone, camera) as they want. Finally, another engineer demonstrates how Microsoft will make even its legacy apps accessible via the cloud. It's a shocker. After years of Microsoft insisting that the desktop is the only proper place for its crown-jewel applications—the venerable Office suite—it appears that Word, Excel, and PowerPoint will levitate from the desktop and become services as well. In this demo, an Excel spreadsheet is running in the cloud with almost all its functionality intact, including features like auto-complete and auto-formatting as well as built-in collaboration and a way to link the spreadsheet results to emails and Web pages.
Editor's update: Microsoft's Red Dog cloud computing platform was formally announced by Ozzie, and officially given the name Windows Azure, at a developer's conference in October. At the same event, Microsoft also provided more details about its online versions of the Office suite. You can read our ongoing coverage of Microsoft announcements on Monkey Bites.
The applause is enthusiastic. The stuff is cool and seems to work. But it's not a thunderous ovation—the new direction is scary. The vision that Ozzie is describing negates the whole question of whether "I'm a PC" is a good thing or a bad one. This future isn't about the PC. It's about the cloud.
"I think we're going to take a lot of people by surprise," Ozzie tells them. That's his low-key way of saying, We're back, baby—and this stuff puts the lie to those who say we're irrelevant. Then his pep talk takes an unexpected turn. "Our greatest challenge may lie within," he says. Throughout its history, Microsoft has demonized competitors—regardless of whether they posed vital threats to the company—and then defined itself in opposition to the presumed enemy. Now Ozzie urges his troops to innovate toward the light, not against the darkness. "Every day we make a choice to focus on the outside competitor or the competitor within," he says, clearly implying that the latter option is the path best taken.
Not long after Ray Ozzie arrived at Microsoft in 2005, he wrote a memo declaring that the company's survival hinged on a shift to cloud computing. Three years later, Ozzie and Microsoft are finally announcing (though for the most part not yet releasing) key products designed to fulfill that promise.
Read more on Monkey Bites: PDC 2008 — More Nitty Gritty Details on Azure
1
Windows Azure Microsoft's long-awaited "operating system for the cloud" doesn't run on a laptop—it runs on the company's thousands of servers. Customers develop their Web-based businesses to operate on Microsoft's data centers, and Windows Azure allocates resources as needed.
Expected launch date: Late 2009
The competition: Google App Engine and Amazon EC2; already available
2
"Zurich" Codename for Azure Services Platform, a set of sophisticated tools to help developers manage their own cloud-based services and Web apps.
Availability Now (in limited preview)
The competition Free open source tools
3
Live Mesh A service built on Red Dog that allows people (PC and Mac users) to synchronize all their files, photos, and music with all their devices.
Expected launch date *2009; now in public beta *
The competition Apple MobileMe
4
Office Web Apps The next major Office release will include relatively complete Web versions of Microsoft's crown jewels—Excel, PowerPoint, Word. Users can subscribe or access free versions supported by ads.
Expected launch date 2010; some apps may appear earlier
The competition Google Docs, Yahoo Zimbra, and Zoho—all available now
Read more on Monkey Bites: PDC 2008: Look Out Google Docs, Here Comes MS Office for the Web
You seldom heard this kind of inspirational, character-building talk from Gates, who stuck to algorithms and product demos in his keynotes. But the original chief software architect is off curing malaria and trying on shoes with Jerry Seinfeld. The current CSA sees his role differently. Enigma or not, Ozzie is the one who must lead—or drag, if need be—a software giant with 90,000 employees, $60 billion in revenue, and an untold number of blue screens of death across a chasm. Can he do it? Ozzie's big advantage is that he knows what's on the other side. In fact, he caught a glimpse of it 35 years ago and has been heading there ever since.
In a sense, Ray Ozzie's remaking of Microsoft began in 1973, when he was a student at the University of Illinois. Ozzie had always been a techie. Growing up in Chicago, he certainly noticed the entrepreneurial push of his dad, a schmoozing insurance broker who cofounded a small agency. But the son was an introvert who gravitated to the privacy and empowerment of transforming bundles of wires, capacitors, and resistors into working gadgets. Ray's little brother, Jack—now a key engineer in Microsoft's software-plus-services push—remembers soldering irons and a good portion of the Heathkit catalog scattered about. "He was like Mr. Electronic Tinkerer," Jack says.
Much of Ray's first two years at the Urbana-Champaign campus was frustrating. He spent endless hours submitting punchcards to the technicians who guarded access to the university's mainframe computer. But on his walk across campus he would pass the old CERL building, which housed the Computer-based Education Research Laboratory. It was a plain-looking five-story brick structure with darkened windows, the activities inside hinted at only by strange orange lights. One day, in the spring of his sophomore year, curiosity drew him in, and what he found there shaped the rest of his life.
"There were rows of orange-glowing terminals and all these people sitting at them," he says. "It was fascinating, because these were plasma-graphics terminals with audio devices, and lots of them." One CERL veteran later wrote about the "large, dull brown cubes about 18 inches on a side, and backbreakingly heavy. What they lacked in appearance, however, they made up for in performance. The display screen, a plasma panel rather than a video tube, showed everything in tiny orange plasma dots ... Every key press was processed right away and response was—usually—immediate."
Ozzie had discovered Plato, an acronym for programmed logic for automated teaching operations. Though largely forgotten now, it was a system almost absurdly ahead of its time, one of those hidden pockets of innovation that bred a cult of followers who even now will bend your ear with tales of their glory days. Only this time, the tales are true. Concocted by a university engineering professor with the weirdly appropriate name of Don Bitzer, the system comprised several hundred terminals, both on and off campus, connected to a CDC 6400 computer. The goal was "automated learning." But what made Plato irresistible to Ozzie was its interactive nature. Users had direct contact and direct feedback—not just to computers but to one another. "They had this thing called Personal Notes, which you would call email," Ozzie says. "They had this thing called Talkomatic, which is like real-time group chat. And they had this thing called Term-Talk, which was like instant messaging." It was also a way-before-its-time Valhalla of computer gaming. Programmers on the system had gone far beyond the tic-tac-toe and hangman that were popular in other computer centers to pioneer multiplayer online games, notably the Star Trek-inspired Empire. In retrospect, looking at the Plato community was like peeking through a wormhole and viewing the 21st-century Internet—but without the spam, fraud, and commercialism that would come with the real thing 35 years later.
In his classic book Computer Lib/Dream Machines, Ted Nelson describes "a seething community of dozens of smart people working like blazes on the project." Ozzie was desperate to be one of them. "I just had to see what was under the hood," he says. "I begged, begged, begged for a job and was eventually hired at like $1.75 an hour." It became the hub of his campus life. "All of us would have gotten higher grades if we had spent less time there," says Len Kawell, a fellow undergraduate who also worked on Plato. "It was such a fun place—a 24/7 party atmosphere, with people writing applications, playing games, sending email, all sorts of stuff. And Ray was definitely one of the hot programmers."
One incident in particular introduced Ozzie to the magic that comes when people connect via computer. He had taken a part-time assignment helping a professor finish writing some courseware. The prof lived on the other side of town, so Ozzie collaborated with him remotely. Ozzie came to know and like his boss, save for one annoyance. "He was the worst typist ever," Ozzie says. "He was very eloquent on email, but on Term Talk it was just dit-dit-dit, sometimes an error, but agonizingly slow." At the end of the project, the man threw a party at his house, and Ozzie discovered the reason for the typing problem: The professor was a quadriplegic and had been entering text by holding a stick in his teeth and poking it at the keyboard. Ozzie was floored.
Plato terminals at the University of Illinois gave users interactivity. "I remember really questioning my own attitudes," Ozzie says. "I had been communicating with him mind to mind. Technology lets you do that, unprejudiced by what anyone looks like. From that era forward, I just knew I wanted to work on something related to communications and interactive systems."
In his first couple of jobs after graduation, Ozzie distinguished himself as a wizard-level programmer. "There was no question that he was exceptional," says Jonathan Sachs, who hired the new graduate in 1979 at Data General, the Boston-area minicomputer company. Two years later, convinced that personal computers were the future, Ozzie joined Software Arts, the Cambridge-based developer of the groundbreaking VisiCalc. "We were a hot place to work, and he was one of our top programmers," says electronic-spreadsheet coinventor Dan Bricklin.
But what Ozzie really wanted to do was reconstruct the Plato world for personal computers. Specifically, his idea was to build an entire online communication system that expanded on a feature called Group Notes, one of the first online conferencing tools—it was sort of a super bulletin board that allowed comments and even conversation on documents posted by users. He saw his chance when he met Mitch Kapor, cofounder of Lotus Development. Kapor loved disruptive ideas and was open to Ozzie's plan, but he had a particular need—an encore for his flagship product, the spreadsheet program Lotus 1-2-3. In 1983, Ozzie and Kapor cut a handshake deal: Ozzie would put together a team to build a suite of applications called Symphony if Lotus would fund his dream project upon completion.
Nine months later, after Ozzie finished Symphony, Kapor made good on his promise. It was an unusual arrangement that revealed Ozzie's business acumen. He structured the deal so that Notes (his name for the new project) belonged not to Lotus but to a company Ozzie would create called Iris Associates. Lotus would have an exclusive option to distribute the final product.
Ozzie (circled) is pictured at a 1982 Software Arts product launch. One of Ozzie's key decisions was picking the operating system on which Notes would run. After the Mac came out in 1984, he knew that Notes had to use the kind of graphical user interface popularized by Apple. But Ozzie wasn't sure which of several competing GUI systems to choose. One option was Windows, but Microsoft had barely announced the first version, and there was no concrete evidence that it would blossom into a viable product. On a trip to Seattle in December 1984, Ozzie and Gates had a long conversation at the Microsoft office—then a single building in suburban Bellevue, next door to a Burgermaster drive-in. Gates had known Ozzie since Software Arts and could relate to him "geek to geek," Ozzie says. He spun a convincing argument for Windows. "You could just tell his commitment to make this thing happen," Ozzie says.
Notes shipped in 1989, but it wasn't an easy sell—it was tricky to install, took some brainpower to grasp, and cost $62,500 for 200 user licenses. "We never had problems believing this was going to be the next big thing, because we used it every day," says former Iris engineer George Moromisato (who now works with Ozzie at Microsoft). "We lived it. We loved it. Ray had already structured his life in a way that today we'd recognize, but back then it was so foreign. His wife and kids were all on email—they had a shared calendar on Lotus Notes, they had access to all their Notes data basically anywhere. It was a peek at what the world would eventually be, and so all of us believed that vision."
The breakthrough came when Sheldon Laube, then IT director of Price Waterhouse, saw the product. "This is going to change the world," he told his bosses. Price Waterhouse ordered 10,000 licenses. The application became a huge success, and Lotus wound up buying Iris for $94 million in 1994. When IBM bought Lotus a year later for $3.5 billion, there was no doubt that the bulk of its value lay in Notes—and in Ray Ozzie. He ran the IBM Notes division for two years and considers it time well spent. "I was cynical of IBM, but they took over Notes when we had 2 million users, and now there are 150 million. With Lotus that never would have happened."
Why cloud computing is everyone's favorite trend.
Tech gurus, venture capitalists, and even the dimmest blog readers know that the digital action these days lies in the "cloud." But what does that mean? Nothing, according to Oracle CEO Larry Ellison. "I can't think of anything that isn't cloud computing," he has said. "It's complete gibberish." But it must mean something—why else would Dell Computer try to trademark the term (as it did in 2007) only to have the request bounced this past summer because the appellation was already ubiquitous?
"The grandiose cover-all version is just software capability delivered over the Net. Consumer cloud computing is the Web," says Dave Girouard, president of Google Enterprise. But when people talk of the brewing cloud battle between his company and competitors like Amazon.com and Microsoft, they're referring to two kinds of businesses. One is operating Internet-based services in the cloud—everything from Web-based programs like Gmail and Hotmail to the model used by enterprise software company Salesforce.com, which hosts and runs high-level business apps for corporations. The other is the idea of offering an Internet-based platform to developers who want to create services but don't have their own cloud to run them on. So they rent storage, computation, and maintenance from someone else.
That last function is the focus of a new software war, to be waged in the cirrus and strata. "There's virtually no limit to the size of this—how big is the whole IT industry?" says Adam Selipsky, a VP at Amazon Web Services. The competitors are the digital titans with their own clouds—dozens of server-packed data centers—and the computer-science PhDs who can manage them efficiently. Downtime is deadly.
Cloud companies assume that consumers will embrace the idea that much of what was once crunched on their PCs and stored on their hard drives will now live in some vague, faraway ether, trusting that it will always be there when they need it. As cloud computing becomes more of a reality, though, skeptics are pushing back. Richard Stallman, the open source guru who founded the Free Software Foundation, has called the trend "worse than stupidity; it's a marketing hype campaign." As Stallman sees it, cloud computing takes away the user's information and locks it into a cloud company's proprietary software system. In his ominous forecast, the cloud is considerably darker. —Steven Levy
Ozzie left IBM and founded a startup called Groove Networks, which made collaborative software. Released in 2001, the Groove app was terrific technology, with peer-to-peer transmission and superstrong crypto built in. But the postbubble timing was awful, and Ozzie realized that the company couldn't make it on its own.
The obvious move was to sell to Microsoft, which had already invested some $50 million in Groove. For Gates and CEO Steve Ballmer, however, getting the technology was just a bonus; the real treasure was its founder. Gates had once described Ozzie as "one of the top five programmers in the universe." Former Groove employees still talk about the time Gates visited and the two leaders got off on a tangent about some arcane technical point. As they bounced improvisations off each other, Ozzie coming up with ideas and Gates rocking back and forth with excitement, it was like watching some propellerhead version of a John Coltrane-Miles Davis performance. Ozzie wouldn't be just a great hire—he would be the hire, the one person qualified to be a partner to Gates and Ballmer in revivifying Microsoft.
But what was in it for Ozzie? Though he respected the intelligence of Microsoft's leaders, he had been repelled by some of the company's hardball tactics. In addition, Ozzie was a lifelong skeptic when it came to big corporations. Plus, he was physically rooted in the Boston area, where he had married a former Digital Equipment Corporation engineer and raised two kids. But Ozzie felt that after losing its antitrust case, Microsoft had tempered its bullying behavior. ("This is a different company," he now says. "It doesn't feel evil; it doesn't feel inconsistent with my core beliefs.") And as for his family life, his son was almost out of college, and his daughter was just entering it, so he thought he could split his time between coasts. The deciding factor, though, was the same one that IBM had offered: impact. Ballmer puts the argument this way: "This is one of the brilliant guys of our time. Now he's got a bigger stage, given by Microsoft, and a bigger transformation, driven by the Internet. It's the ultimate opportunity for one of the world's great drivers, thinkers, and technical minds."
The first months at Microsoft weren't easy. Ozzie soaked in the campus culture, gently sharing his big ideas for the next stage of technology. "People were telling me, 'Go evangelize—inspire people and the company will transform.' But after six months of walking around the campus, I couldn't see how we were going to get from point A to point B." Microsoft was focused on getting its big products, a revamped Office and the deeply troubled Vista, out the door. "It wasn't thinking about what to do next."
He decided to write out the thoughts he'd been brewing, as Gates had done a decade earlier with his famous Paul Revere-ish Internet memo. On October 30, 2005, Ozzie's manifesto, "The Internet Services Disruption," was emailed to about 100 top execs and managers. (It later went out to everyone.) The sender wasn't Ozzie, though, but Gates, who told recipients it would be "as critical as the Internet Tidal Wave memo was when it came out."
The Ozzie memo was an explicit call to action, as well as a subtle but stern criticism of the company's current state. In his description of the technological landscape, and despite language designed to cushion the blow, he told of a corporation in danger of missing the boat:
Our products have embraced the Internet in many amazing ways. (Implied message: not enough.)
But for all our great progress, our efforts have not always led to the degree that perhaps they could have. (We're falling behind.)
And while we continue to make good progress on many of these fronts, a set of very strong and determined competitors is laser-focused on Internet services and service-enabled software. (And we're not.)
Even beyond our large competitors, tremendous software-and-services activity is occurring within startups and at the grassroots level. (Twentysomethings in garages are surpassing us!)
Many startups treat the "raw" Internet as their platform. (While we're still trying to finish Vista!)
Ozzie's top lieutenants at Microsoft's Windows Live Core offices are (from left) David Treadwell, Debra Chrapaty, John Shewchuk, Jack Ozzie, and Amitabh Srivastava.
Photo: Lionel Deluy Microsoft, Ozzie wrote, had to think and operate more like an Internet company and, as much as possible, like a Web startup. Consider ad-supported or subscription business models, he advised, viral distribution, and experiences that "just work." Instead of the clunkiness that Microsoft products so often displayed, focus on being "seamless." Bottom line: Change big-time, or else.
"We were clearly marching along some of those directions, but we were taking an incremental approach," says John Shewchuk, a Microsoft technical fellow who would become an Ozzie lieutenant. "Ray pulled it together in a comprehensive vision."
"It was a big and important memo," says Ballmer, who feels that the Ozzie dispatch called for even tougher adjustments than Gates did in 1995. "The Internet didn't require a change in business practices, just technology changes," Ballmer says. "The notion of moving toward more subscription-based models, more ad-based models, is a bigger change for more people."
The memo didn't get immediate results. Basically, Microsoft couldn't consider a paradigm transplant until Vista and Office shipped. Ozzie was undaunted. "I kept talking," he says. "I began incubating certain things, certain new projects off to the side." While a number of Gates' duties were passed to others (notably to chief research and strategy officer Craig Mundie, who was happy to make public appearances as Microsoft's tech ambassador), Ozzie took on some of Gates' product reviews and strategy functions while quietly building a team to implement the ideas in his memo. That team included some key personnel from Ozzie's previous enterprises. "The reason I'm here is Ray," George Moromisato says. "I just believe in his vision." Jack Ozzie, Ray's brother and Groove cofounder, also came along.
Getting the best and brightest at Microsoft to leave their sinecures and join Ozzie's team was more of a challenge, but Ballmer's active involvement helped land the bigger fish. "I'm not going to break your arm to do this job," he told VP David Treadwell. "But I'm going to twist it pretty hard." Treadwell spared himself injury by becoming a key member of Ozzie's operation, heading up a team called Windows Live Core.
Ozzie spent a lot of time crafting a different kind of work environment at Microsoft. "He was very intentional about getting stuff done quickly, focusing on the end customer," Treadwell says. Previously, a big part of any development team at Microsoft was making sure its new product worked in lockstep with everything else the company produced. This "unification" criterion was something that Gates had always hammered on. But Ozzie saw that while that approach avoided annoying conflicts, it also tended to smother innovation in the cradle. "This philosophy of independent innovation—really making progress before you pursue serious integration, is something Ray pushed very strongly," Treadwell says. Ozzie's approach was to encourage people to rush ahead and build things. Then he'd have a team of what he calls the spacklers fill in the gaps and get things ready for release.
In a sense, his teams were cultural pioneers modeling a more flexible, startup style of software development. As a signal of his new approach, Ozzie spent a lot of time on the physical workspace for his team. "I'm either at 1 inch or 30,000 feet," he says. "If there's something I care about, I'm on it." He had workers rip down the labyrinthine, catacomb-like corridors on one floor of a building on the Red West campus and called in architects to create a more open design. Now, walking into the Windows Live Core group is like leaving Microsoft and visiting a Futurama set. Office windows open onto hallways so that quick eye contact can trigger spontaneous discussions. Whiteboards are everywhere. Pool tables, mini-lounges, and snack zones draw people toward the center of the space rather than isolating them around the edges.
At first, the skunk works-like nature of Ozzie's operation engendered suspicion and resentment. "There was a perception of 'Who are these special guys off in their own little space working for this new guy?' That created some tension," Treadwell says. By breaking seldom-questioned rules of how space should be apportioned in the Redmond Borg, Ozzie was pushing against the battleship mentality. Management experts might appreciate his explanation of exactly who was putting up the resistance. "The company," he says. Who's that? "They. You know—Microsoft."
Ozzie acted less like an executive dispatching his duties per the org chart and more like a startup CEO in touch with every aspect of the process. "At first, people go, 'Whoa, what is this—are you a product manager? Are you a developer?'" Jack Ozzie says. "Then they warm up, and man, they just want to be part of the team he's leading." That's fine for the groups he spends time with, but will it scale? "None of that can happen overnight," Jack says. "It's going to have to spread from the people who work for him." Also essential, Jack adds, is whether Ozzie's teams actually deliver: "We're a very results-oriented culture here."
The Ozzie project that must deliver results is Microsoft's so-called operating system for the cloud. As more apps become Web-based, the raison d'être for Windows—running programs on desktop PCs—becomes less compelling. What better way to make up for the decreasing importance of a desktop operating system than to create a dominant OS that runs services in the cloud? This is not only a crucial effort but one in which Microsoft is playing catch-up: Amazon.com went live with its cloud services in early 2006 and now hosts data storage or applications for more than 400,000 developers, including the complete historical archives of The New York Times. Google's entire company is based on the premise that people want to move from desktop to cumulus. But Microsoft hopes to use its cloud OS (codenamed Red Dog, now called Windows Azure) to dominate the cloud the way DOS and Windows did the desktop.
Ozzie found his project lead for Red Dog in Microsoft veteran Amitabh Srivastava, a top computer scientist who had been pulled from Microsoft Research to fix the engineering process for the troubled Vista. Their first meeting, at Ballmer's urging, was set for 4:45, which implied a hard stop after an hour or so. "One of the rules in my family is that nobody misses dinner," Srivastava says. But when Ozzie came in, the two chattered like magpies, and sometime after 8 Srivastava realized that he had indeed missed the evening meal.
Srivastava agreed to head Ozzie's Cloud OS project once Vista was done. His first coup was getting Dave Cutler to come out of semiretirement to join him on the project. At age 66, Cutler is a legendary figure in computing. He wrote the groundbreaking VMS operating system for DEC and later led the team that created Windows NT. An irascible, impatient combination of John von Neumann and Sgt. Rock, his presence on the team gave it instant credibility. Srivastava and Cutler began by methodically visiting every Microsoft group working on services, from Xbox Live to Virtual Earth. In December 2006, Srivastava wrote his own vision document outlining the plan. Its title? "Owning Clouds."
Under Ozzie, Srivastava felt free to create Red Dog using methods not normally seen at Microsoft. He set up his own 1,000-machine data center right in the middle of the Redmond campus to test early versions. To power the operation, the team stole excess reserve power from three nearby buildings. No permissions were sought through Ballmer or Gates. "I take direction from only one person—Ray," Srivastava says. Another indication of the rebel nature of the project comes from its codename. "The official story is that we are just like Red Dog beer, and I'm sticking with that," Srivastava says. But Cutler is more forthcoming: "We were visiting Hotmail," he says, "and there was a really seedy strip joint in San Jose called the Pink Poodle. I said, 'Maybe we ought to name this project the Pink Poodle.' Everybody said, 'Oh, God, we could never do that.' And then somebody said, 'Red Dog,' and we all said, 'What a great name.'" Cutler also had the idea of outfitting team members with shiny red Nike sneakers.
Maybe the most subversive aspect of Microsoft's newest operating system is that it was produced with a fraction of the manpower the company usually directs to critical projects. "There are literally thousands of people on Windows, but small groups with very focused people is a better way of doing things," Cutler says. "So this project is much smaller. It's like 150."
Red Dog, available late next year, will have competition, of course, from Amazon and most certainly Google, whose own cloud OS, App Engine, will offer developers similar hosting benefits at lower cost or even for nothing. But Debra Chrapaty, the Microsoft exec in charge of the company's data centers, says that Microsoft's infrastructure is so efficient it can compete in cost even with a company she refers to by the letter G. (She refuses to speak its name out loud because "every time you say that word, it reinforces their brand," she says.)
Eric Schmidt, CEO of that G-word company, says that because Microsoft has so much market share in servers and operating systems, the Redmondites will certainly be big players in cloud computing. He sees it as an extension of Microsoft's nasty behavior in the '90s. "Microsoft's basic strategy is to gain enough share in cloud computing to force other people to use its standards," he says. (By contrast, Google has blessed an open source version of its cloud technology, which both IBM and Yahoo have adopted.) Ozzie doesn't buy the charge. "Google and Microsoft have the same basic philosophy. We're basing our cloud on Windows technologies because they're great technologies and we have a lot of higher-level services on them. If you want to write open source stuff on them, you can do that."
Srivastava and Cutler predict that Red Dog's reliability will be a competitive edge. "We don't want to say to developers, 'Hey, come and use this platform,' and then have it lose your data," Cutler says. "That would just be bad. I mean, it would be terrible. So we're being really conservative."
Ozzie, of course, loves Red Dog and prizes his red sneakers. But the project closest to his heart may be Live Mesh, one of the building- block services that run on top of the cloud OS. In a sense, it addresses a problem that then-VP Jeff Raikes posed to Ozzie during his first week at Redmond: How can Microsoft connect and synchronize people's contacts, calendars, and other information in a seamless fashion?
For Ozzie, this challenge was profound; it meant using the cloud to connect people to machines and, more important, to each other. "Getting sync right is the essence of everything," he says. "If you don't, everything else fails." And furthermore, it had to be designed in a way to scale to hundreds of millions of concurrent users. "Scalable sync is tough." Apple knows this all too well; when it turned on a similar service called MobileMe last June, it delivered more stink than sync.
Live Mesh looks impressive in demos; it can zip photos and tunes from computers to the cloud. Then it can zing the information off to anything that's connectable, smoothly fitting the stuff into its natural habitat. Whether it will work with millions of people and billions of items—and so simply that those millions of people keep using it—is another matter. (Microsoft's "it just works" record is spotty, to put it gently.) To ensure success, Ozzie has pared down the project to essentials. David Treadwell says the team jettisoned some bells and whistles. "But Ray was very firm," he adds, "about maintaining the soul of the release."
Brothers Jack and Ray Ozzie hang out in 1979. To Ozzie, software's soul does not lie in the accumulation of features. Instead, it lies in his dream of connectivity. "Live Mesh is very Ray," Mitch Kapor says. "It's the son of Groove, which is the son of Notes." Which was, of course, the son of Ozzie's beloved Plato. Thirty-three years later, Ozzie is still trying to build on what he saw in sophomore year. But it's no longer the Ray Ozzie vision. It's Microsoft's.
"Here's the deal," Ozzie says. "Somewhere in my first year, not in the first few months but before the CSA announcement, I had to make an internal decision. If I want to be here, is it to make Microsoft successful or to have a good project, a good experience, whatever? It was an issue of engagement. I asked myself, why do I do what I do? I enjoy solving complex problems that involve technology, people, organizations—the whole mix. So I made the internal decision to do what I can to make Microsoft successful, and that was it. Yes, it took me a while to understand what that meant. No, I haven't worked here for 20 years. But every day I'm up at 5 am and at work at 6 or 6:30. I don't get home until 8. I'm doing everything I can to make this company successful."
That just means getting Microsoft's 90,000 people to follow him into the clouds.
Senior writer Steven Levy (steven_levy@wired.com) wrote about the creation of Google's new Web browser in issue 16.10.
Inside Chrome: The Secret Project to Crush IE and Remake the Web
Nov. 10, 1983: Gates Opens Windows a Bit Early
15th Anniversary: Name-Dropping 'Gates' and 'Jobs' Again, Again ... and Again
The Many (Geeky) Faces of Bill Gates
