As I have pointed out before, without real adoption data, it is hard to gauge whether customers are really interested in a particular platform. We know server adoption data for Linux with reasonable accuracy. At last glance it is large, and rapidly growing, outpacing the market growth, and its rivals growth. This tends to have secondary effects, on to those in a moment.
We do not know desktop usage growth across the market in any detail. That said, we have deployed multiple Linux desktop workstations for users as their primary system. We see a strong growth in it, and are hopefully well positioned to deal with demand (multicore units with gobs of processing power and ram). Note: We tried this recently with windows x64 (which I liked) but the lack of drivers for critical bits for our customers wound up preventing them from using it. Every (open source) driver in the kernel in Linux available for 32 bit systems is available for 64 bit systems, with pathological exceptions (hard wired 32 bit codes).
One of the side effects of platform growth, real platform growth that is, is that developers ramp up targeting it. They will target a platform they believe they can make money on. They will leave platforms they cannot. Unix has lost (or is losing) most of its developers in favor of Linux. Now it appears that this is occurring in windows as well.
Now back to that pesky adoption data. Way back here I had postulated that adoption rate data was critical. Moreover I noted that without it, if the market adoption rate for CCS was 0.1% or so, that
Of course, it could be a 0.1% (19 clusters per year of about 32 nodes in size) in which case it is a $280k market to Microsoft.
Now count how many clusters were on that Microsoft press release from ISC. Lets assume, generously, that 1/10th of the clusters installed with CCS made it onto that list (I personally guess it is closer to 1/2 to 1/3), so lets call it ballpark about 150 clusters with CCS (remember, I am being generous). This puts the entire CCS revenue below $28M. Again, I make assumptions on cluster size.
Like above on the market targeting, we want to know how large the windows HPC market *really* is. It is (and has been for a while) fairly obvious that the Linux market is growing at the expense of some other market(s). On the HPC side, Linux has been driving this market, with staggering growth rates. In order for the CCS to be compelling, it needs to exceed these growth rates. Sustainably.
I am not convinced it is. Moreover, there is mounting evidence that Linux adoption is growing at a faster rate than the market is growing. That market share has got to come from somewhere. OSX isn’t growing appreciably (though iPhone is about to double/treble the installed base overnight) as far as I am aware (and please point out to the contrary if this is true). Solaris is exhibiting minimal growth, slower than the market. Thats sort of like a bank account with an interest rate below inflation. Yeah, the amount of money is increasing over time in it. But the value of that money is decreasing faster than the rate at which the amount of money is increasing … net you are losing money. Solaris is doing that with relative market share.
Developers are going to follow what they perceive to be the growth markets. Why invest money when there is no return? Similarly, why develop for platforms upon which the return is low when you can invest that same effort for markets where the return is high … . Developers have to be pragmatic.
If you want to know what a company is going to be doing in the next 3 to 6 months, look a who they are hiring (or firing). If you want to know where the future apps are going to be, ask the developers. They look at adoption data. Some do trailblazing. Most have to see an “R” higher than the “I” to justify the effort. More importantly, adding platforms increases costs. Those costs need to be offset by future revenue and profits. Increasing platform count is not a good thing unless the platform you are expanding to is growing rapidly. And even then, you want keep the costs down, so if you adopt a new one, you might be dropping an old one, where growth is stagnant, and there are few new or existing customers.
Users follow developers, and developers follow users. Its a vicious cycle. Chicken and egg.