[CS-FSLUG] Quantify the Power Savings (was: A Cost Analysis of Windows Vista Content Protection)

[Alan Trick]

In certain cases this actually is correct. When we're talking about
Total Cost of Ownership, the statistics droids are looking at one thing
only: what is more cost effective *right now*.

Obviously the result is going to be misguided. After all, retraining
your staff is a fact of life. But the reality is that a lot of people
are more effective with Windows at the present than they would be with
Linux.

Porting applications is another mess to get over. The Win32 API is a
mess, and once your application has gotten tangled up in it, it's pretty
difficult to get out. And for Enterprise-level customers, Wine does not
cut it.

As for which is more efficient on the same hardware. Linux is almost
always better off here, but hardware is cheap, retraining people and
getting software developers to port your legacy applications is much
more expensive.

Alan Trick

On Fri, 2006-12-29 at 16:33 -0800, Greg Slade wrote:
> I was reading my usual trade rags today, complete with the Microsoft 
> propagand^H^H^H^H^H^H^H^H^Hads claiming that Windows has a total cost of 
> ownership than Linux. (It still boggles my mind that nobody has sued M$ 
> over that little bit of book-cookery.) There was also much discussion 
> over power usage in data centres. According to one article, 38% of all 
> power consumed in the US is burned up in data centres (both in powering 
> the servers, and in cooling them), with all the associated costs to 
> businesses and the environment. Taking those two together, it wonders me 
> if anybody has ever done a study quantifying the electricity savings with 
> using Linux vs. Windows. After all, we all know that Windows is bloated, 
> and wastes far too many clock cycles in OS overhead. What I'd like to see 
> is some lab results showing, say, useful (as opposed to theoretical) 
> MFLOPS using the same hardware running Linux vs. Windows to perform the 
> same tasks. I'd expect it to show that running Linux allows you to get 
> more work done on the same hardware for the same amount of power used, 
> but has anybody actually run the tests to put hard numbers on the 
> difference?
> 
> Greg
>