Okay, if I hear another one of these studies by folks who have nothing to do but count useless pieces of information, I am going to scream! Once again, we are inundated with various posts discussing the down turn in the Linux server market. But, once again, how do these PHD’s with a calculator and pencil count?

The server market used to be an easy thing to count. If you bought any kind of Intel system, you were going to run Windows variants of some sort. If you purchased a Power series from IBM, you were going to run AIX. And, if you bought any Sun hardware, you were going to run Solaris. So the bean counters really had it easy; no real thinking required.

Fast forward to today - where can I run Linux? I can run Linux on Intel based systems. I can run Linux on Power series hardware. IBM even will give you a “canine and equestrian” of the virtues of Linux on pSeries! I can even run some distributions on Sun hardware - although don’t know why I would… And then, there is the mainframe. I can even run Linux on the mainframe.

So, what happens when a server is purchased? Most companies, small to large, purchase in similar fashions. They have a purchase agreement or contract with a vendor of the hardware - in this case we are talking xSeries exclusively - and with that agreement comes a brand new license for Windows server software. Only recently, a large company that I do work for, finally figured out that they should not purchase the license from the hardware vendor. What’s even more interesting is that larger companies usually have some type of “global distribution” license agreement with Microsoft to distribute Windows server licenses. So, if hardware is purchased by large company ABC, does the bean counter assume Windows is going on it?

I think the more interesting numbers would be, what kind of server is being purchased, and what is it’s use. Many companies go through a cyclical “end of life” exercise regarding their hardware. In many cases, if a company is serious about saving money, they will consolidate. If the consolidation platform is xSeries, that means it is running some form of virtualization software. Then, how do the bean counters count the O/S to hardware purchase ratio? And what O/S is assumed to be loaded?

So all that to say this; I still find this discussion to be a complete waste of time. Who are we trying to impress or make feel better about Windows having an advantage? I would hope they would! They’ve only had the advantage in the market for just over a decade. I mean, they should have taken over the world by now!

My advice to the counters - start counting something that matters! Those who are firmly entrenched in Linux serverdom have no interest in your numbers, and those firmly entrenched in Windows serverdom are just nodding like the sheep they are. At the end of the day, most independent consultants responsible for sound consolidation implementations and Linux opportunities are stacking the deck quietly on the data center floors.

Spread The Word:
  • Digg
  • Sphinn
  • del.icio.us
  • Facebook
  • Mixx
  • Google
  • blinkbits
  • blogmarks
  • Fark
  • Furl
  • Ma.gnolia
  • NewsVine
  • Reddit
  • SphereIt
  • StumbleUpon
  • Technorati
  • TwitThis
  • YahooMyWeb
  • LinkedIn
  • Webride
Sphere: Related Content

One Response to “Linux Server Market Down the Tubes?”

Kiana Tom

November 15th, 2007 - 2:55 am

Hey!…Thanks for the nice read, keep up the interesting posts..what a nice Thursday

Leave a Reply