Over the course of the past two decades, Sun Microsystems has been up and down a number of times.
The first time Sun was down was when the Unix workstation market was supposed to disappear due to the emergence of the Intel workstation. But Sun moved into the server market and was then followed into that space by Intel servers. With the coming of the Internet boom, Sun managed to stave off Intel servers in a world that valued performance. But with the collapse of the Internet economy, the heat on Sun became too intense, and the company experienced a general meltdown.
Since then, Sun has been trying to right itself by emphasizing software under the leadership of company president Jonathan Schwartz. But that effort has met with only limited success and, at its core, Sun remains very much a hardware company.
So the question is, what will Sun focus on to regain its competitive edge in hardware?
The man at the heart of that question is John Fowler, executive vice president for Sun's Network Systems Group. Fowler, along with Andy Bechtolsheim, Sun's chief architect and senior vice president for Network Systems, is spearheading a project within the company that is sometimes referred to as Galaxy.
According to Fowler, Galaxy initially will be a family of systems based on AMD processors that will run Solaris, Linux and Windows under a common systems management scheme. This may not seem like something all that special, but it actually represents a seam that Sun can exploit between IBM and Hewlett-Packard.
IBM enjoys a great reputation when it comes to server hardware, such as the xSeries. But its systems management software, provided by its Tivoli software unit, has always been an afterthought. HP, meanwhile, has some of the best server management software around, but it is usually outclassed by IBM hardware.
What Sun is going to try to do is deliver rock-solid hardware coupled with first-class, self-healing systems management software as part of a bid to deliver a combined offering that has the best ROI in town. On top of that, Sun will also be the only vendor in the class solely focused on AMD processors--in contrast to IBM and HP that sell both AMD and Intel, or Dell, which is focused solely on Intel. Later versions of the Galaxy family of products will also feature Sparc processors.
Fowler and Bechtolsheim won't stop there. Sun is also working on its next blade server designs, due out in 2006. Fowler knows that Sun is comparatively late to market, but he also notes that blade volume remains small compared with the rest of the market, which he said means that Sun has plenty of time to catch up.
A lot of people think Sun's current financial condition means that the company won't have the financial resources required to reinvent and successfully market a completely new set of hardware products. But while Sun's success is far from guaranteed, it's worth noting that the state of the art in server management software is still relatively rudimentary compared to what could be done.
All one has to do is look at the sophistication of the tools available in Unix and mainframe data centers to get an inkling of the tools missing from the x86 server space. And it shouldn't surprise any one if Sun winds up delivering the next generation of self-managing x86 servers before anybody else. They have the expertise and, perhaps more importantly, they need it more than anybody else.