On Sat, 10 Nov 2001, Leonard, Robert B Mr EACH wrote: > Here's a link re: the amazon/intel migration stories. > > Let's keep the FUD in redmond. > > http://www.wininformant.com/Articles/Index.cfm?ArticleID=23086 Well... I wouldn't exactly call the wininformant article "unbiased." There is some subtle use of FUD in his article by mixing statements like, "lack of robust office packages" with an article that should be only about server OS migration. What's this got to do with servers? Who knows, but some people might interpret it as, "*Linux* isn't ready so don't use it." The purpose of FUD is to create Fear, Uncertainty, and Doubt in the mind of the customer so they instead stick with the "safe choice" and don't choose a competitor's product. By making Linux look like there's no real office suite (apparently the only qualifying "real office suite" is MS Office), he paints Linux as a whole as somehow unsuitable for business use. He also quotes the Intel rep saying that the overall cost of Windows and Linux is almost identical, once you factor in support and maintenence. Two problems here: (1) Are we just referring to a single Windows install vs. a single Linux install? Usually companies have multiple Windows installations which means the cost of Windows is going to go up with the number of installations because each additional server/workstation adds the cost of an additional license. (2) "Almost identical" could be interpreted as $100 difference or $100,000 difference, depending on the person's perspective. This perspective is most obvious when the OS cost is factored in with the hardware cost, where it's usually seen as insiginificant in comparison. He also downplays Linux's importance because it *only* has 24% of the server market (according to the IDC figures he quoted) and that Windows has 38%. To me, when I hear that Linux has 24%, I'm astounded. Just a couple years ago most people hadn't even heard of Linux and now it has 24%. That's pretty impressive growth in a very short period of time, although I will concede that this rapid growth will slow down because most of that growth has been at the expense of proprietary *nix boxes. If you've got a business that has been running only Windows for the last 5 years and only has Windows admins, well, then that company probably isn't going to be switching to Linux anytime soon. These admins would have to spend time outside of work teaching themselves Linux to a level where they felt comfortable deploying it on the network. I'm guessing that there's a very small percentage of Windows admins that have the time/energy/motivation to do this. If Microsoft's licensing changes really pissed them off then all they'd have to say is, "We won't upgrade anymore." Or, if they're a big enough company, I'm sure they could use the threat of switching to Linux as a negotiating tool against Microsoft for a better licensing arrangement, whether or not they actually would switch. So future server growth for Linux is probably going to slow somewhat, but with a lot of college CS/EE students picking it up, it will probably continue to chip away at Windows server installations as these individuals enter the job market. One of the common fears that managers have about deploying Linux is that there's not a large enough pool of qualified Linux administrators available. So they decide to stick with Windows because the belief is that there's always a Windows administrator available. This is true, for the most part, but I'm guessing that it will change in the near future. Also, if you read his comment in the user posting area, he basically wrote this article to combat a lot of the excitement surrounding Linux and to make it clear that Amazon didn't switch from Windows to Linux; they switched from *nix to Linux. But the end result is the same: A large and visible company has shown that Linux has what it takes to run their business. As far as I can tell, the majority of Amazon's cost savings came from switching to cheaper x86 hardware, but this also means they chose not to use Windows on that platform. Most likely it was too costly to rewrite their back-end code for Windows, but it's still a win for Linux and a huge promotion for it's future success, in my book. ~M