Interesting view of how Linux would have worked out, had Sun done the right things with x86 processors.
But this strikes me as denying Sun's position, as per Clayton Christensen's "Innovator's Dilemma". Sun passed on a number of opportunities for disruptive innovation. For example, the folks who started Auspex came from Sun, I think (and in turn, Auspex missed the boat that took Network Appliances away). Sun had a processor that was pretty darn good, when intel released the 80386. Let's face it, SPARC architecture still makes a lot more sense than x86. And Sun kept up for a surprisingly long time. How many people did Intel have squirrelling away at the x86 architecture? How many people did Sun have on SPARC? I imagine it was 10x less.
I think suggesting that Solaris could have taken the place of linux in the market is very much wishful thinking. Now, if Sun had listened to Larry McVoy's Sourceware Operating System Proposal in '93 (http://www.bitmover.com/lm/papers/srcos.html), that might have made a big difference.
I agree. Sun was dependent on selling proprietary hardware at $20k a pop. Even if they had launched a commodity-hardware line, no one would have paid the prices necessary to support a corporate structure built on high-end sales. SGI tried, and it didn't go so well for them.
Sun didn't even need an expensive salesforce to sell them - we were a government research customer in the 90s and had all our orders for Ultra sparcs cancelled - Wall St were paying above list price for any that Sun could build
Assuming that google would be running on solaris "if only we had..." strikes me as being a bit myopic. I think it's more likely they'd be running on freebsd.
Sun never committed to the low-end. Their sales droids wanted to sell to the enterprise and big .edu .
Of course, now Linux is what enterprise and big .edu are running... I am told that Solaris has a tenacious hold on certain financial shops and banking, and other niches as well.
It is a shame, their $995-list 1U servers (V100, Netra X1) use only 70W including powering 2 hard drives. Had they kept up and released a follow-on series with SATA and more RAM expandability, they would be perfect for cloud applications.
But this strikes me as denying Sun's position, as per Clayton Christensen's "Innovator's Dilemma". Sun passed on a number of opportunities for disruptive innovation. For example, the folks who started Auspex came from Sun, I think (and in turn, Auspex missed the boat that took Network Appliances away). Sun had a processor that was pretty darn good, when intel released the 80386. Let's face it, SPARC architecture still makes a lot more sense than x86. And Sun kept up for a surprisingly long time. How many people did Intel have squirrelling away at the x86 architecture? How many people did Sun have on SPARC? I imagine it was 10x less.
I think suggesting that Solaris could have taken the place of linux in the market is very much wishful thinking. Now, if Sun had listened to Larry McVoy's Sourceware Operating System Proposal in '93 (http://www.bitmover.com/lm/papers/srcos.html), that might have made a big difference.