Analysis: Where did Sun go wrong?

Just before the crash of the dot-com boom that spawned the meteoric rise of Sun Microsystems Inc., then-CEO Scott McNealy and President Ed Zander held a meeting where they discussed the future of their company.

At the time, Sun owed the fast-growing sales of its server hardware and position as an industry darling to the dot-com economy, “an unnatural period when they were considered a 'you must have this vendor' company,” said Illuminata analyst Jonathan Eunice, recalling the meeting he attended in 2001.

Even as the boom that had allowed their business to flourish was about to end, the two executives did not appear to see anything amiss, Eunice said.

“It was a moment when everyone in the room knew that things weren't going to continue as good as they had been,” he said. “But they were on stage saying, 'We don't see anything falling down.'” It was a “moment of blindness” for Sun, he said, one that came at a turning point when the company should have been making tough business decisions to ensure its longevity.

The years since that meeting in 2001 have not been kind to Sun. The company that had built up its business selling powerful Unix servers faced a dual threat in the form of x86 processors, which would spawn cheap but powerful servers that undermined its Sparc-based offerings, and the Linux operating system, which would be embraced by IBM and Oracle Corp. and emerge as a potent, low-cost alternative to Sun's Solaris.

Those events were beyond Sun's control, of course, but while rivals like Hewlett-Packard Co. and IBM reinforced their businesses by embracing the new technologies and building up their professional services arms, Sun's reaction was to hunker down. What followed, analysts and even some former Sun executives say, was a series of missteps that have left the company where it is today — it has lost $1.9 billion over the last two quarters, with revenue down 9% to $6.2 billion. Indeed, the company has had difficulty sustaining profitability since the dot-com bust. Moreover, negotiations to be acquired by IBM have apparently broken down, at least for the moment, leaving Sun in a weak and exposed position.

It has been an ignominious and somewhat mystifying downfall for a company that, for all its missteps, has been one of the greatest innovators in Silicon Valley history. From Java to Jini to utility computing and its slogan “the network is the computer,” Sun has often shown an uncanny vision for identifying industry trends before they happened. In many instances it was, quite literally, ahead of its time.

“It's weird because Sun is probably the most innovative and forward-looking company of any that I've seen,” said John Crupi, CTO at enterprise mashups company JackBe and a former CTO for Sun's Enterprise Web Services group. “They were really good at doing cool engineering things. The problem was, they never managed to move up the stack and package those things into solutions that really let you use the technology and solved people's problems.”

Stephen Hultquist, an IT industry veteran and principal with Infinite Summit in Boulder, Colo., agreed that Sun never suffered from a lack of innovation and vision. Where its executives faltered was in their ability to get beyond the vision and give customers what they wanted to run their business, he said.

“Any company that doesn't continuously focus on the real benefits to their customers — both current and future — will eventually lose the mindshare of those customers,” Hultquist said. “Customers buy what makes sense for their success and satisfaction, not what is 'best' in some esoteric, technical way. If that wasn't the case, Microsoft would not be the powerhouse that it is.”

For instance, “the network is the computer” was Sun's slogan long before companies like Google Inc. and Salesforce.com Inc. built their businesses around the idea, and before cloud computing became the buzzword du jour. But now many of the companies embracing cloud computing and hosted services run Linux rather than Solaris in their data centers.

Sun also was one of the first to see that subscription-based software pricing would replace the traditional per-CPU model when multicore processors became prevalent. But by the time Sun had got itself together to offer software in this model, there was less interest in its Java software because they were already using BEA or IBM products.

Marie Beeson, who worked at Sun from 1999 to 2004 in its professional services division, compared Sun to Xerox PARC in its ability to create useful technologies but failure to capitalize on them. Xerox PARC created a lot of the technologies that made computers and the widespread use of the Internet possible — such as the GUI and Ethernet — but is a research facility and not a profitable business.

“From a technical standpoint, Sun was very innovative and had a lot of great technology, but it didn't know how to really exploit those technologies,” she said.

Observers identified three principal errors that, had they been avoided, might have allowed Sun to flourish after its dot-com business dried up. They are: not reacting to Linux by open-sourcing Solaris more quickly than it did, not building an x86 product line fast enough to sell alongside its Sparc systems, and not learning how to monetize the great Java technology it created.

Sun finally came out in support of Linux in 2003, around the same time it launched its first x86 server line. The next year it open-sourced Solaris through a project called OpenSolaris. But by then Linux already had the support of IBM and Oracle, which eased the concerns of customers worried about big-company support behind the open-source operating system, and lower-cost x86 boxes running Linux or Windows were replacing Solaris in many enterprises.

Some feel that if Sun had open-sourced Solaris sooner, Linux would not have become the popular low-cost alternative to Unix that it is today.

Joe Lindsay, A vice president at engineering for interactive media firm Brand Affinity Technologies, thinks Sun could have headed off the Linux phenomenon by focusing earlier on its software business, rather than favoring its hardware, from which it derived most of its revenue.

“If they had focused on becoming a software company like Microsoft — selling OSs and software on top of that — Linux may have never needed to be created,” he said. “If Solaris was as accessible and available [on servers] as Windows on the desktop is, then Linux may never have happened.”

Lindsay left a job at IBM in 1995 to work at companies with Sun-centric data centers because at the time he was a big fan of Solaris. Now he works mainly with Linux and other open-source software that Sun did not create, even though Sun has now open-sourced almost all of its products, including Solaris, that once were proprietary.

For sure, some say Sun could have done nothing to stop the march of Linux, even if Sun could have made moves to compete more effectively against it. Mitesh Tahker, a consultant at Laksh Systems, which serves the New York financial services industry — once a major market for Solaris but now a Linux stronghold — was a Linux early adopter, and said the OS took off because it was vendor-neutral, not just because it was open source.

Failing to turn Java into a large, profitable business early on — which IBM and others did so successfully — was the third of Sun's tactical missteps. Though the company created Java as a development technology in 1995, it didn't really put its application-server house in order definitively until early 2004, when it pulled the components of its Java software stack together under the Java Enterprise System (JES).

By that time, IBM and BEA (now part of Oracle) were already the only two Java-based software vendors that mattered, and the open-source JBoss app server — now part of Red Hat — was on its way to commoditizing the Java middleware stack.

Sun did have an application server called iPlanet long before JES. But it took too long to make iPlanet compatible with the then-J2EE standard, now called Java EE — an important criteria for some customers, and one that Sun helped to develop. Sun struggled because it had acquired three technologies — NetDynamics, Forte and Kiva — and not one of them was J2EE-compliant, Crupi, the former Sun Web services CTO, said.

“Weblogic was all Java, and it became J2EE-compliant very quickly, while Sun was struggling to integrate all this technology that wasn't J2EE-compliant and they missed the boat.”

Sun's focus on innovation over products started at the executive level, some former employees said. McNealy, who insiders said privately was a more thoughtful, approachable executive than his querulous public persona would imply, managed in a way that fostered an explosion of creativity but wasn't particularly in tune with market dynamics.

Miko Matsumura, who spent five years at Sun as chief Java evangelist, described the environment at Sun in particle-physics terms. Matsumura, now a vice president and deputy CTO at Software AG, said McNealy liked to watch his employees collide intellectually with one another because he thought it would spur new ideas.

“Scott was a very clever manager in the sense he would hold a frame and allow people to compete within that frame,” he said. “He loves to see people smacking into each other. He really wants to see the best quality bubble up and set conditions for things to collide.”

The byproduct of that kind of chaos is that there were too many new ideas — such as what happened to Sun's Java app-server business — and no single clear message about how to turn ideas into a viable business plan, Matsumura said.

When CEO Jonathan Schwartz took over from McNealy in April 2006, he tried to clarify the company's direction and make it more in line with the software and hardware products customers wanted to buy. Under Schwartz, Sun has open-sourced a lot of its valuable proprietary software and hardware assets, giving customers more choice about how and on what platforms they can deploy Sun's copious technology.

Sun spokesman Noel Hartzell said the company has seen a “significant” increase in its software business since it open-sourced Java and purchased open-source database vendor MySQL in a deal that closed early last year. And Sun's Open Storage business, which combines open-source storage software with Sun hardware, also is seeing success, with 21 percent year-over-year growth in the quarter ended Dec. 28, 2008, he said.

Sun also has worked hard to take costs out of its business and improve operational efficiency, saving billions of dollars over the last few years, Hartzell said. Cost savings that included layoffs announced in November should amount to another $900 million this year, he said.

But cutting costs historically has not done enough to boost Sun's profitability. And with the IBM deal apparently off the table and Sun's business lagging overall, long-suffering shareholders likely won't wait much longer for Sun to bring them a return on their investment. Despite executives' best intentions, Sun's efforts may be too little, too late.

Many suspect that Schwartz will be pressured at some point to leave his post, much like Yahoo Inc.'s Jerry Yang after he walked away from a deal with Microsoft last year. Some rumors even see McNealy returning to his old job to attempt a turnaround. If he does, the pressure will be on for him to do for Sun now what he was unable to accomplish back in 2001.