Sun shines dimly in Big Blue's shadow

Does Sun Microsystems have what it takes to remake itself in IBM's image?

Sun Microsystems has been in the news again. Predictably, its most recent earnings report disappointed shareholders and left analysts and customers alike wondering, again, just how the company plans to reverse its flagging fortunes. Given the current economic downturn, the prognosis seems plenty gloomy.

Myself, I can't help but compare Sun to another well-known enterprise IT vendor: IBM. In many ways, IBM is the perfect model of success for Sun to emulate. The problem is that Sun may have already traveled too far down the wrong road to reinvent itself in Big Blue's image.

Is open source enough?
Lately Sun has cast itself as, among other things, a leading open source software company. It has spent the last few years open-sourcing the Java platform, and earlier this year it purchased open source database vendor MySQL for $1 billion. Its stated aim is to make its entire software portfolio available as open source, backed by subscription service and support offerings.

But pundits are divided on the effects that a prolonged economic slump might have on the open source software business. On the one hand, author Andrew Keen asserts, "Mass unemployment and a deep economic recession comprise the most effective antidote to the utopian ideals of open-source radicals." In other words, developers are likely to forsake open source as they scurry off to whatever paying jobs they can find. On the other hand, new Red Hat CEO Jim Whitehurst expects his company to "pick up quite a bit of new business where companies are looking to save money from what they are doing."

So where does this leave Sun? Superficially its software business model sounds just like Red Hat's, yet a closer look reveals something a little different. Sun's new top investor says that it is "progressively less of a server company and more of a software company." But Rich Sands, Sun's community marketing manager for Java SE, revealed an alternative assessment of the company's software strategy in a reply to Sun executive Simon Phipps's blog last week:

The real Sun model ... is that ubiquitous software platforms driven by potent open source development and deployment communities creates "opportunities to monetize" hardware and services -- aka "drag." You still need racks of humming machines to run all that free stuff. Sun's challenge: deliver hardware and professional services that run the free stuff with better ROI than anyone else's gear.

This, in a nutshell, is the age-old business of being a "solutions provider." When push comes to shove, risk-averse enterprise customers are usually more than willing to buy an expensive piece of hardware with all the necessary software pre-installed -- provided it comes with a service and support contract -- rather than cobble the same system together themselves from cheap parts. Assurance trumps cost savings.

Battling the budget blues
But how soundly? Sun has always played to the top end of the market -- IT haute couture, if you will -- and that's going to be a tough sale given current market conditions. In a blog post last week, Sun CEO Jonathan Schwartz admitted that the company was "surprised by the magnitude of the decline" in revenues from its traditional, high-end server products during its first quarter of 2009, which Schwartz attributed to the economic downturn.

It will be bad news for Sun if this downward trend continues. While Sun has enjoyed some success with its line of x64-based servers, as a company it really isn't geared to fight it out with the likes of Dell and HP for the low end of the server market -- particularly when the mastermind of its x64 successes is on his way out. More importantly, that kind of server business won't support a software business based on open source.

In a reply to another commenter on his blog, Phipps remarked that Sun's software efforts "are definitely not intended as a loss leader," and indeed Sun's revenue from its software business has grown year over year. Yet by its own statements (PDF) Sun's software revenue amounts to just one-tenth of what it earns from servers and other systems and one-fourth of what it earns from storage products.

Just how intertwined Sun's various divisions might be is known only to Sun. But it seems clear that software support subscriptions alone aren't going to keep this beast running -- not by a long shot. Sun's fortunes are inextricably linked to its ability to sell solutions.

And this is where the comparison to IBM comes in. Because while Sun has spent the last several years slowly circling the rim of the whirlpool, IBM has emerged as a colossally successful IT vendor by doing precisely what Sun needs to be doing: selling solutions.

IBM offers servers, big and small. IBM offers software to run on those servers. IBM offers open source. In fact, according to some analysts, the total revenue that IBM brings in from open source alone rivals Red Hat's annual earnings. But IBM's business model differs from Sun's in two important ways.

Where IBM beats Sun
First, although it is a staunch supporter and contributor to open source, where IBM has invested heavily in software, IBM sells software. WebSphere, DB2, AIX -- it cost a lot to develop these products, and IBM has consistently shrugged off any suggestions that it open-source them.

Compare to Sun. Sun likes to position itself against Red Hat, but Sun is no Red Hat. One recent Linux Foundation whitepaper estimates that it would take approximately $10.8 billion in 2008 dollars to build the Fedora 9 Linux distribution, upon which Red Hat Linux is based, from scratch. That's $10.8 billion that Red Hat didn't have to spend, because Red Hat's product is essentially a found resource.

How much has Sun spent to develop Solaris or Java? How much must it continue to invest in maintaining other products, which, despite being open source, have no appreciable development community? To say these products are not loss leaders suggests something akin to Hollywood accounting.

Meanwhile, IBM sells a ton of Linux and has contributed loads of code to the Linux kernel and related packages, but it doesn't maintain its own Linux distro; it leaves that to Novell and Red Hat. Sun has to learn to choose its battles as effectively if it hopes to compete.

The second major difference between IBM and Sun lies in IBM's emphasis on consulting. It's no secret that Global Services saved IBM from the brink of failure in the 1990s. Today, more important than the servers or the software that runs on them, an IBM solution comes complete with the know-how to make it all work smoothly.

Sun spent years putting products first. Only recently has it come round to the services-first model of today's IBM, but it's a model that's yet to be proven effective for Sun. It says a lot when customers would rather buy their Java or their MySQL from IBM.

As Rich Green says, the key will be not just delivering products and professional services, but doing so with better ROI than anyone else. Sun needs to convince its customers not merely that it offers an impressive product and service portfolio, but that it can be as valuable a partner as IBM. Until it can do that, Sun is going to have a hard time crawling out from under Big Blue's shadow.

UPDATE (11/18/08) - The original version of this article attributed a quote to Rich Green, a former Sun exec. The quote was actually from Rich Sands, a current Sun manager. I apologize for the error.

Copyright © 2008 IDG Communications, Inc.

How to choose a low-code development platform