For most people, Sun president and COO Jonathan Schwartz is a somewhat enigmatic figure because he combines hippy-like sensibilities with the analytical skills of an academically trained McKinsey consultant, which is job he once had prior to going to work for Sun when the startup company he founded was acquired by Sun.
Since joining Sun, he has emerged as the cerebral voice that complements Sun CEO Scott McNealy's famous hip-checking approach to the market. Like all consultants, Schwartz is exceptionally good at making a case and there is no doubt that he is smart. But what is in doubt is whether he's right.
During a meeting with editors from publications such as CRN, New York Times, Wall Street Journal, Wired, Business Week and Red Herring, Schwartz went into detail about what he sees as a fundamental shift in the ecosystem of the computing industry. Like most discussions, it was long on vision and short on specifics that will turn that vision into a reality, but an overview of some of Schwartz's views is well worth considering.
On the positive side, Schwartz sees the industry as a whole move to a managed services model based on utility computing. He sees this as a huge opportunity for Sun to regain its lost luster because people running mission-critical managed services will have a much higher regard for the reliability and robustness of Solaris 10 running on low cost AMD or SPARC processors than they do for commodity operating systems running on Intel platforms.
While that may be true in the long term, the question is whether Sun will be viable enough to be considered more than a niche player when the market finally becomes aware of the issues he is talking about. And it also assumes that Linux development will be standing still as opposed to becoming a much richer platform thanks to the contributions of professional developers from IBM, Oracle, SAP and other vendors committed to the platform. Furthermore, he seems to discount IBM's ability to match any type of managed service play Sun cares to deliver.
There's no question that Sun has come up with an aggressive pricing model for its managed services that is requires customers to only pay for the amount of CPU time consumed. And he correctly argues that low server utilization rates and the cost of the electricity required to run those servers means the industry needs a new approach to data center computing that is based on managed services.
But it is naive to argue that IBM can't respond to Sun's pricing challenge if it so chose, even if Sun is committed to relentlessly driving prices downward. IBM could easily respond, but right now there is no real pressure to respond to Sun's challenge. Sun, in terms of competing with IBM Global Services, is little more than a pesky fly flying around the head of a giant. Should HP, EDS, Dell and every other vendor adopt the new Sun pricing paradigm, you might then have enough Lilliputian weight to tie the IBM's Gulliver Services down. Worse yet, Sun makes it seems like telecommunication carriers are going to be the transport mechanism for all these managed services. That may be true by the end of the decade, but right now the companies that provide managed services today are solution providers in the channel, many of whom have no relationship with Sun. If Sun is counting on the carriers to fulfill its vision, then sometime in 2009 some of us may be around to see the world he is describing.
In much the same manner, Schwartz is also tilting at Linux by promising that Solaris will soon be a more viable alternative that is also open sourced based. Should this happen, the approach Sun will take will be more like the open source license of MySQL and Apache, rather than Linux. Of course, Schwartz will argue the real competition for Solaris is not Linux itself but rather the Red Hat implementation of a stack of software on top of a Linux kernel.
But whether an open source version of Solaris is relevant remains to be seen. An open source Solaris is not going to attract legions of developers. And worse yet for Sun, it would juts bring more pressure to open source Java. Sun likes to say it needs to control Java to prevent it from forking under the pressure of multiple open source distributions. But there's no reason why they shouldn't let the market determine what might truly be the best open source distribution in the same way that Red Hat has become the dominant Linux distribution. After all, if Sun has the best testing and certification procedures, they should be able to easily minimize any open source Java distribution. So short of open sourcing Java, the best path for Sun may to not create an open source version of Solaris that at best may be considered irrelevant and at worst considered disingenuous.
The thing that Sun has going for it is that the world craves a highly reliable computing platform that is secure and reliable. Sun has that in the form of Solaris on AMD coupled with a complementary Linux offering. Focusing on value rather than grand statements of vision is what will win the day.
To that end, Sun should consider obsessing about winning back the financial services segment of the market at the expense of every other vertical. The financial services space is a big part of the company's base, but the customers in that space are a known entity that Sun's existing sales force and partners know how to reach. Where Sun is an unknown entity with a compelling value proposition is in the mid-market segment where customers have never been more willing to listen to alternatives in a difficult economic climate.
Oddly enough, two other near and present opportunities that don't seem to be as high on Sun's radar screen as they might is the potential of the Java Desktop System. While it's not likely an open source client supplant Microsoft Office in most environments, there is a need for a low cost client that serves as the foundation for a highly integrated composite application that spans multiple systems. The Java Desktop System coupled with XML Web services is much better suited to play that role than Microsoft Office is given that most of the data sources that matter in these environments are moving towards Java.
And finally, while SPARC systems may not be able to ignite a revolution, they are very compelling options for embedded systems, especially those running Linux. Sun has a huge OEM business but seems blind to the opportunities that custom system builders represent. The fact of the matter is there will be more raw computing horsepower going into systems outside the office and applications such as RFID are only the tip of iceberg.
Sun has always been technically smart and business lucky. The client/server revolution around Unix servers followed by the Dot Com boom made the company what it is. But for Sun to succeed going forward it needs to finally step up and make its own business luck. Right now, Sun just seems to hoping that it can hang on long enough to catch another lucky paradigm shifting break in the form of an emerging utility computing that will take five years or more to play out. And that's a shame, because the real value proposition for Sun could easily be enormous if it only spent more time worrying about the value of the here and now and less about seeing what might be happening three plus years from now.