Hardware losses doomed Sun but Linux hurt too

Sun's hardware losses were the primary cause of its decline and subsequent purchase by Oracle, but lack of foresight in Linux and opensource market along with many Sun missteps contributed to its downfall.

Less than six years after Sun Microsystems' then-CEO Scott McNealy sneered at Linux as a "hobbyist," build-your-own-jalopy solution, it appears that the mighty Penguin is going to beat the formerly $14 billion Santa Clara, Calif.-based colossus to the finish line. And the much-trashed prediction of Linux Foundation Executive Director Jim Zemlin last fall that "the future is Linux and Microsoft, not Unix or Solaris," appears to be a...

race-winning bet.

Now that money-losing Sun has gratefully agreed to be swallowed by Redwood Shores, Calif.-based Oracle Corp. for $7.4 billion this summer, it's fair to ask: Did Linux cause Sun's downfall?

Not if you look at the numbers. After all, Sun's rapidly growing software revenues are only about $600 million a year. Hardly enough to topple an enterprise still raking in some $12 billion in annual revenues. Sun's fortunes were staked on its high-performance, overpriced hardware, not its software, analysts agree.

It was the hardware that hurt Sun, ultimately.
Mark Driver
Research Vice PresidentGartner, Inc.

Indirectly, however, it's a different story. The huge success of Linux, which Sun's competitors pitched in to develop, created an operating system (OS) alternative that eroded demand for Sun's pricey systems for basic computing needs, and boosted sales of Sun's rivals' lower-cost x86-based machines.

"It was the hardware that hurt Sun, ultimately," said Mark Driver, research vice president of Stamford, Conn.-based Gartner Inc. "The reality is that Intel and others killed off Sun. I think Sun's fortunes were already set by the time they aggressively leveraged open source."

Sun's next-generation hardware sales also were hit harder by the 2001 dot-com bust than rivals like Hewlett-Packard Inc. and IBM Corp. because Sun's biggest market was the Web-driven e-commerce sites, Driver said. Sun's competitors, on the other hand, were entrenched in traditional data centers and weren't as affected by the downturn, he said.

More on Sun and Linux
Following Oracle-Sun agreement, open source worries persist

Open cource community reacts to Sun-Oracle deal

Has the Sun set on Sun Microsystems? Analysts say no

Sun CEO: Linux is for 'hobbyists,' not enterprise

Sun's mistakes
In many ways, however, Sun was its own undoing. Interviewees cite Sun's superior, isolationist attitude, its lack of business strategy and inability to capitalize on its assets like Java, as well as its lack of insight into key industry trends like Linux and low-cost x86 commodity hardware.

"If Sun had been more aggressive in developing an Intel-based server line instead of competing with Intel [with its Sparc servers], we might have a very different market altogether," Driver said. "By the time Sun launched openSolaris [for x86 chips], IBM and HP had been in that business for years and had a significant advantage and Linux was the de facto Unix flavor."

Sun also mistakenly thought that software was simply a feature of hardware and, except for Java, slighted any software project unless it drove hardware sales, he said. Although Java was successful, Sun allowed competitors like IBM and Oracle to reap its financial rewards, he added.

Finally, unlike its hardware rivals, Sun failed to adapt fast enough to the rising popularity of open source, Driver said. This mistake, which Microsoft also made, was especially costly to Sun because a transition from Unix to Linux is relatively easy, he said.

By the time Sun fully committed to open source and, more recently, buttress its portfolio with MySQL in an effort to bleed revenues from its competitors, Sun had missed the market and was losing too much money to offset its hardware losses, Driver added.

The Linux Factor
But according to Ted Ts'o, CTO of the Linux Foundation, competition from Linux "has everything to do" with Sun's fiscal downturn.

Competition from Linux "has everything to do" with Sun's fiscal downturn.
,

Linux gave customers a choice of software vendors, cheaper hardware and a large developer community thanks to the unprecedented collaboration of many corporations including IBM, Hewlett-Packard and Intel Corp., Ts'o said. When Sun finally adopted open source, not only was it late in the game but its go-it-alone effort left Sun footing Solaris' entire development cost, with a much smaller community base, he said. In addition, Sun failed to recognize that open source OSes would become a commodity, with revenues solely generated from additional software services, he said.

"When Sun switched from a proprietary to an open source model, they didn't change the fundamental business dynamic and that's what sunk them," Ts'o said.

Making matters worse, Sun's delayed expansion into the x86 market came at the expense of its own Sparc processor sales which, in turn, reduced its revenues, according to Al Gillen, program vice president, system software, with Framingham, Mass.-based International Data Corp.

As for the future, Ts'o said Sun's hardware rivals have been successfully coaxing Sun's hardware customers to convert to x86-based Linux systems "for a number of years," a trend he predicts may accelerate with the uncertainty about the platform following the Oracle acquisition due to be completed in August.

Ts'o questioned how much of Sun's pricey hardware product line Oracle will be able to maintain without hurting profit margins, especially with Sun's post purchase drop in earnings and continuing deficits. He also expressed concern about how many Sun software engineers Oracle can afford to retain.

"Solaris and MySQL will probably continue, but as a shadow of their former selves," Ts'o said. "But Sun has hundreds of other, lesser known open source projects… that may be in peril" if Oracle halts the work without releasing the code to the open source community, he said.

Ts'o said Solaris will continue in some form but he doesn't know if it will become a "zombie program loader" for an Oracle appliance or continue as a full OS. For clues, look for a release date for Solaris 11, supposedly due out in 2010 (the last full version was in 2005), and watch to see if Oracle promotes x86-based Solaris systems by investing in new ISV relationships, Ts'o suggested.

"If Sun is going to be a money-maker, Oracle will have to cut back, and I predict a lot of Sun employees will be out of work in six months," he added.

While it's "far too early for predictions," Driver doubts that Oracle would reduce Solaris to a program loader in the near future because Solaris is too much of a "cash cow," he said. As for future adoption, Driver says that some users, especially Oracle shops, may be happier with Oracle support and a single vendor instead of two. But others running SAP or DB2 may opt to switch.

Innovation following the merger may be a "mixed bag," but the cultures of the two Silicon Valley companies should blend well if Oracle can retain Sun's infrastructure and staff, he said.

"This acquisition propels Oracle into the center of the open source universe," Driver said. "Oracle will leverage that position for all its worth."

Let us know what you think about the story; email Pam Derringer, News Contributor .

Dig deeper on Unix-to-Linux migration

0 comments

Oldest 

Forgot Password?

No problem! Submit your e-mail address below. We'll send you an email containing your password.

Your password has been sent to:

-ADS BY GOOGLE

SearchDataCenter

SearchServerVirtualization

SearchCloudComputing

SearchEnterpriseDesktop

Close