An article on softpedia discusses the state of the windows/linux client and server markets. I can personally attest to seeing Microsoft scrambling to try to turn every Linux opportunity into a windows win, but I am not sure they completely grasp what they are up against. This article amplifies this.
Customers will continually increase to take advantage of the inexistent price tag that is associated with Linux. IDC forecast that free Linux server operating system deployments will outpace new subscriptions in the next four years. Nonpaid Linux server deployments will jump to 43.3% from just 41.4% in 2007. “This phenomenon is not unique to Linux ??? as we see the same trend playing out with Windows server operating environments,” Gillen noted. ???We find that more customers are seeing nonpaid Linux as a viable solution for certain non-critical business needs, despite the lack of commercial applications and the potential support challenges that come with a non-commercially-supported distribution.”
While I take issue with some of their wording … many customers we know are using Linux, free or paid subscription … precisely for the critical business needs, due to the control and cost issues.
Your Nth license of Linux costs you nothing (thanks to the GPL). If you want to pay for support, you can. You can pay Redhat, Novell, Canonical, and others. You can pay firms like mine for support. You are free to choose the support you need, and what you are willing to pay for. You do not have to pre-pay for support, but then you need to be prepared to pay per incident, which will be higher.
Redhat’s business model is selling support subscriptions. And it works, fairly well. I have a number of issues with their kernel, and some of the choices they have made, but otherwise, they have a fine system … its just not that good for HPC. Great for web servers, fine for other functions. SELinux, 4k stacks, and no xfs? Yeah, this isn’t so great for HPC.
Microsoft is conflating Linux with Redhat. I’ve had some exchanges recently with some staff that suggests that they don’t yet grok the concept of free installation, not requiring any payment. They also don’t seem to grok how easy it (generally) is to install. We have it fairly well automated (not as much as I like … it works well for RH 5.x, SuSE 10.x/11.x, and we are working on the Ubuntu 8.04 version). So our time spent loading it is quite low.
The OS is a detail, and should be driven by choice of application and its requirements. If you can lower the total cost of the system by selecting a different component that has few negative consequences associated with the choice (such that you come out at a win relative to the original choice), in this day and age, it would be hard to justify making the choice for this higher cost system.
If all OS choices cost zero, then select the one you want if the application will work on it. If one of the choices costs $X more than the rest, does it either bring more than $X benefit back to you? That is, is the ROI more than the additional cost? In the cases of OS, its really … really hard to make that argument.
Linux does cost $0. Its the support that costs money. This is the business model. Don’t call up Redhat for a 1024 core problem if you have 1 license, they will not likely help. You need to pay them for the support you need. This is the open source model. And its why Redhat is growing so fast and hard.
I should also point out that many of our customers are in fact noting that when deploying desktops, we are hearing that Linux will likely be used. In part to keep the same OS on the desktop as on the cluster, and in part to keep the client costs down.
Remember, Linux costs nothing to deploy. Zero. Zippo. Zilch. You can install it on a million machines, and pay no company for this installation.
CIOs, and everyone below them, are now quite cognizant of this.