Yes Solaris has been around for some time. But there is a significant SunOS base out there, and there will be for some time. On a second front there is the issue that UltraSPARC's reported performance is only achieved on V9 SPARC code. There is a large amount of V8 and even v7 SPARC code that is still out in the installed base, that will run as fast or faster on the new HyperSTATIONs than it will on an UltraSPARC machine.
Academic computing environments are significantly different than corporate environments. First off, corporations usually are not getting the large discounts on the hardware that Sun will offer schools. Secondly, academia basically get's the software for free, or close to it much of the time. Hence, a license for a new V9 piece of software doesn't factor in since it's free. Most of the time, schools probably should choose UltraSPARC because they get signficant discounts on the hardware, and upgrading their software doesn't cost them anything. Academia is often on the cutting edge because they can afford to be. I saw this in grad school where we had every single piece of Mentor Graphics latest software installed, and tons of Sparcstation 20's. The software was free, and the computers were very cheap. (this was 3 years ago so the sparcstation 20 was top of the line). This was literally millions of dollars worth of hardware and software that the school got for thousands.
Corporations can't work that way. Ross offers customer's with installed bases a large array of options. Upgrades are just one of these options. Hyperstations, motherboard upgrades to Sparcstation 5's, the new SparcPlug systems fill out a pretty broad product line and hit a bunch of price/performance points.
I think Ross can win many of the technical arguments of why people should buy their products. I believe what will determine the success of the company in the next couple of years is how well Ross can communicate this message to the SPARC market. Right now it looks like it's Ross's sales/marketing team vs. Sun's team. Sun's going to try and convince people that V9 SPARC is worth switching to today, Ross is going to argue that staying with V8 is cheaper and offers the same performance on your software base and that customers should wait until the V9 market matures a bit before upgrading. Who's right? I guess the market will tell us over the next year or so.
On the long term front, Ross is also moving to V9, it's not like the company plans on staying in the 32-bit SPARC market forever.
-Sean |
