I'm in a small, daylight filled conference room in Sun Microsystems Menlo Park campus with John Loiacono, head of Sun's software group. We've just met and sat down but he springs back up and walks over to the whiteboard, "let me just show you something . . ."
This is normally not a good sign--whiteboards are just one step away from PowerPoints--but more squiggly. I usually try to avoid either and prefer a conversation rather than an explanation of every rectangular box or graph in a PowerPoint--or whiteboard doodle.
But Mr Loiacono is already drawing half-a-dozen rectangular boxes and I decide to quell my normal yelps of protest. And in this case, I'm glad I did because the whiteboard did its job--Mr Loiacono was able to draw me a picture of Sun's business strategy and explain it in such a way that makes sense.
For too long, Sun has been making announcements of products and acquisitions in a piece-meal manner--and far too many of them. It made the Sun story too confusing.
Now I see Sun's business strategy, and the single most important thing about a business strategy is that a company has one (you'd be surprised at how many don't). The second most important thing about a business strategy is that it can be communicated simply to customers, staff, business partners, and investors.
Sun still needs to work a bit on the second part--let's see if you can do it without a whiteboard. But overall, Sun has a strategy, it has a story, and it has some momentum--which is a good thing. A couple of miles down the street is rival Hewlett-Packard, which doesn't seem to have much of a business strategy--or at least they haven't dragged me in yet to show me their whiteboard.
The Sun strategy is very focused on grid computing. Sun knows that grid computing is the most efficient way to deliver computing as a utility and that that is the future. How soon we all get there, is a story for individual companies and their requirements.
"Some organizations will want to own their own grid, others will be happy to utilize part of someone else's grid, and we will be the supplier to the grid builders," says Mr Loiacono.
Open source software is a key element of the Sun strategy. In fact, the entire operating system and middleware, plus development and management tools is open. Which means Sun's Solaris is open source too. But Mr Loiacono laments that few people realize that Solaris is open source and has far more features and capabilities--and users--than Red Hat Linux.
It is probably because Solaris has such a long legacy of being a proprietary operating system, I suggest. And that is something which will take years to dispel. In fact, Sun might be better off in renaming Solaris as Sun Linux--at least that would get the message across that it is an open system, and more robust for enterprise applications than Red Hat or Suse.
Branding issues aside, Sun is making all the software in its middleware stack available for "free" and charging a monthly fee per user for support, upgrades etc. "Most enterprises don't know how much their software is costing them because of all the different pricing models. We give them one price and so they know their total costs and can better plan for new projects."
Sun, wisely, doesn't want to be in the applications space, but it does remain firmly in the hardware space. However, it has shifted some of its focus on systems development and is happy to use Advanced Micro Devices Opteron chips in its blade servers, and also to use high-end SPARC microprocessor designs developed by long-time partner Fujitsu.
Sun also realized that in a future world where servers in the data centers will be measured in the hundreds of thousands--electric power consumption becomes a key cost. Power is needed to run and cool the systems and that's why low-power consuming chips such as Opteron and SPARC microprocessors are essential.
But birthing a brave new world of grid computing requires Sun to set an example, and so it operates a commercial service that provides grid computing services to enterprises. "We don't want to be in the grid computing services business we want to be a supplier, but we do want to show the viability of the business."
Sun expects that the telcos of the world will be the ones that will be operating massive grids and they have the sales force that will sell those services to enterprises. Sun wants to be a vendor to those companies.
And this is where I think there is a potential problem for Sun. Because the telcos of the world are not the best business partners. They are slow moving, regulated, and they are not motivated to reduce costs as quickly as other companies, in more open markets.
Sun might consider building grids that are customized for different industries--in addition to being a vendor to other grid builders. Because it knows how to run and tweak the IT systems it builds better than anybody else. It can build and sell entire grids to service providers, or it can continue running them as services while others run the sales force.
This would accelerate the competitive landscape in the grid services market, which would lead to more sales for Sun, and cheaper machine cycles for everyone else.
But will the use of grid computing, and it's resultant greater utilization of computing resources across entire sectors, mean less need for quite so many servers in the future?
"No, I don't believe it will," says Mr Loiacono. "History shows us that if you make more computing power available people will use it. And they will build applications that will fuel the need for more computing."