Migrating from Windows to GNU/Linux

Can you save money by moving your business to GNU/Linux? The short answer is that you can. However, the long answer is that how much you save — or if you save at all — depends upon your resources and choices. If you are lucky, you might find a study comparable to your situation to help you plan, but most of these studies are biased one way or the other, so you should still need to do your own assessment as you plan the move.

Whether a switch is worth your effort is usually calculated in terms of Total Cost of Ownership (TCO). As the term emphasizes, TCO includes not only the cost of buying and installing the software, but also the cost of hardware to run it, administering it, training employees to use it, and maintaining and supporting it. Sometimes, TCO is divided into first year costs and yearly costs, to reflect the extra expenses of setup and configuration, and the presumably lower costs of running an ongoing system. But, however TCO is calculated, what matters is that making the switch is not just as simple as assuming that, because GNU/Linux has no licensing fees, you will automatically save money.

The truth is, the TCO of a switch can vary wildly, depending on your expectations and planning. To give just a few examples:

  • Do you have the in-house expertise for the migration, or will you rely on consultants?
  • Are you switching just your servers, or employees’ desktops as well?
  • Will you get technical support from on-line forums, or will you be more comfortable with a traditional support contract from a company like Red Hat or Novell?
  • Will you take advantage of GNU/Linux’s lower memory requirements to coax another year or two of life out of your old equipment, or buy completely new hardware?
  • Do your employees need only minimal adjustments to switch from Microsoft Office to OpenOffice.org, or will you want them to take a complete training course?
  • Will you need new employees to maintain your new infrastructure, or can existing ones cope, possibly with retraining? And these are just some of the considerations. To complicate matters even more, your assumptions will also affect your projected TCO. To give just one example, how should you estimate the cost of hiring a GNU/Linux administrator?According to Datamation, in 2009, a system administrator’s salary ranged from $52,750 – $82,500 in the United States, to which an extra 10% could be added for GNU/Linux expertise — which implies that the expertise is still relatively rare. At first, you might conclude that the migration will involve increased administration costs if you have to make new hires. However, many analysts argue that, given the general reliability of GNU/Linux, your new administrator will have more free time, and be more productive than a Windows administrator, and therefore save you money.
  • Similarly, as you compare the costs of Windows and GNU/Linux, should you assume that both operating systems have the same costs for administration and maintenance? At first, you might think that a reasonable assumption.However, experienced administrators almost unanimously report that GNU/Linux is more secure, more stable, less affected by viruses and malware, and easier to patch and upgrade, as detailed in “Open Source Software vs. Commercial Software: Migration from Windows to Linux.” If you accept this testimony (and it is comprehensive enough to seem convincing), then assuming roughly equal maintenance costs will falsify your estimate, skewing it in favor of Windows.Multiply these examples by twenty or thirty, and you will begin to understand the difficulties of estimating TCO. Even a simple task such as estimating the cost of obtaining licenses for Windows or Microsoft Office can be difficult because of ever-changing policies. Con Zymaris of Cybersource, an Australian company that has tracked migration costs for almost a decade, notes that “it has become increasingly difficult to obtain clear figures for Microsoft licence and software assurance costs. One has to be an expert or be within the Microsoft licencing fraternity to make heads or tails of it.” The more you look into each factor, the more complex it can be.


Faced with the difficulties of estimating TCO by yourself, you might be tempted to search online for an estimate that you can apply to your own circumstances. However, be warned that, as with many contentious issues, TCO analysts show a strong tendency to conclude what they already assume.

This tendency is especially strong among those whose conclusions favor Windows. In one notorious study for the Yankee Group, Laura DiDio concluded that, while GNU/Linux might save money for small companies, generally the total cost of switching was three to four times more expensive and took three times as long as upgrading from one version of Windows to another.

However DiDio’s conclusions not only run counter to most other studies, but contain numerous assumptions that betray an unfamiliarity with GNU/Linux (such as a failure to factor into its relative immunity from viruses when estimating maintenance costs). DiDio was also criticized for lumping enterprise and small and medium businesses together and ignoring their differences, and for other aspects of her methodology. The report was quickly debunked, although some Microsoft sites continue to cite it as evidence.

Other unbalanced pro-Windows assessments are included in Microsoft’s Get the Facts campaign, which ran from 2004-2007, and continues to run on Microsoft sites outside the United States, such as Microsoft Canada. The Get the Facts site itself offers numerous marketing case studies full of sweeping generalities and unsupported claims.

For instance, the case studies for C.R.I.S. Camera Services states that”By switching from Linux to Microsoft software, C.R.I.S. is improving its business efficiency while at the same time reducing the effort required to manage the technology that supports the business. Microsoft Dynamics™ NAV gives the company a powerful business management solution that has improved technician productivity by 20 percent and will help the company to realize an expected 40-percent increase in annual revenues.” These kinds of statements lack the detail to any use whatsoever, except as clumsy propaganda.

Needless to say, pro-GNU/Linux studies should be approached with equal caution. However, what distinguishes the most convincing pro-GNU/Linux studies is their careful explanation of how they arrive at their figures, which allows you to study and question the conclusions.

One example of these studies is Cybersource’s 2004 study of GNU/Linux TCO. The study carefully explains its methodology, in particular pointing out that it has made several choices in Windows’ favor, including ignoring the cost of fighting malware and viruses and of unexpected downtime and tripling the estimates of costs for external GNU/Linux consultants. In addition, the study summarizes and critiques other TCO studies. Even more importantly, the study gives each cost as a line item, so that you are left in doubt about how the conclusions were reached (and can copy the line items for yourself).

Cybersource’s conclusion? In 2004, a company of 250 would save 36% by using a non-commercial version GNU/Linux with existing hardware, or 26% if new hardware was purchased. If a commercial version of GNU/Linux such as Red Hat or Novell were used with existing hardware, the savings would be 26%, while, with new hardware, the savings would be 19%.

Cybersource’s figures are six years old, and more recent TCO studies are hard to come by, especially ones that lists every line item. Possibly savings would be a few percentage points lower today, because Microsoft and its partners have lowered their prices in recent years in an effort to become more competitive.

Still, here and there, some indicators emerge. For example, the French Gendarmerie Nationale reports that, by switching to Ubuntu, it has reduced its IT budget by 70% without reducing its capabilities, and saved fifty million Euros in licensing fees in five years. Similarly, the same is true of the German Foreign Ministries’ claim that, by using GNU/Linux, it has reduced its annual costs per desktop from three thousand Euros to one thousand — by far the lowest of any German ministry’s.

Admittedly, these indicators are not detailed. But unlike the claims on Microsoft’s Get the Fact site, they are at least first-hand reports from those involved in the migration. Even allowing for a wide margin of error, they suggest that, with the right choices, switching to GNU/Linux remains an effective cost-savings measure.

By contrast, you have to search long and hard to find pro-Windows studies that are either detailed about their assumptions and methodologies or else first-hand reports. In fact, I have been unable to find any, although conceivably they exist.


David H. Wheeler has argued in detail why you should at least consider GNU/Linux. However, as explained to Robin Miller in an interview for Linux.com in 2004, “in the end, the only way to be really sure that you have unbiased results is to do the comparison yourself — which you have to do anyway, because some measures like total cost of ownership (TCO) and performance are incredibly sensitive to specific environments.”

Wheeler recommends looking at other reports as you develop your own TCO estimates. However, he also suggests that you look at how the report was funded, so that you can assess its contents in more depth.

“Even potentially biased reports can give you some useful data, as long as you’re careful with them,” Wheeler says. “A report paid to review a vendor’s own product will often raise issues that vendor thinks are to the vendor’s advantage — but those issues might be very important to you, and thus worth thinking about (and examining the competitor for that attribute). Also, these vendor-sponsored papers often identify who that vendor thinks is valid competition — so make sure you include that other vendor in your evaluation! For example, Microsoft has information comparing OpenOffice.org to Microsoft Office (previously noted in Slashdot). So as an acquirer, that’s a tip-off that if I’m thinking of buying/upgrading Microsoft Office, I’d better also consider OpenOffice.org.”

Such research takes time, to say nothing of an alert and analytical mind. In many cases, I suspect that an open-minded TCO assessment will lead you to considering GNU/Linux, but there may be some factors, such as a lack of trained staff or key employees unwilling to learn new software that make you cautious.

However, the point is to make the decision for yourself. Switching infrastructure can be a huge disruption to your business, so you owe it to yourself and your business to evaluate the choices openly before you begin.


4 thoughts on “Migrating from Windows to GNU/Linux

  1. apexwm

    Great article. I’ve also posted my own detailed explanation of why businesses and individuals should migrate to GNU/Linux as well, at the URL posted here. Thanks!

  2. Pingback: Can you save money by Switching to a Linux based OS ? | The Linux Journal

  3. GreyGeek

    The Windows TCO angle always ignores one BIG factor: the cost of infection. It is TOO BIG a factor to ignore and can easily be the most expensive part of Windows TCO.

    It is not if a Windows installation will get infected, it is only a matter of when. The fact that Microsoft keeps news of security holes secret until such time, if ever, that they deem it necessary to announce the hole and the patch on the same day, then have the hubris to claim “zero day” patches, renders as moot any claims by Windows users that their installation is not and never has been infected. One cannot protect against holes of they have no knowledge. It reminds me of the “up times wars” of a decade ago, when Windows users were claiming uptimes that matched or exceeded Linux up times claims. That “war” came to a sudden end when Microsoft announced the 49.7 day clock bug affecting ALL Win9x installations. NO computer running any version of Win9x could stay up more than 49.7 days before it would hang or involuntarily reboot. That fact meant that Windows fanbois were lying about up time claims which exceeded 50 days.

    If large corporations with a paid and trained IT staff using the best up to date security software cannot keep their corporate desktops and servers free from infection, what chance does Joe and Sally Sixpack have?

    Sadly, Joe and Sally find out that Windows is insecure when they get dunned for merchandise they did not buy, some so expensive that the resulting expenses and legal bills bankrupts them. Such items have included houses and cars, purchased by bad buys with their personal credit information.

    Even consumers who don’t own or use computers but were doing business with TJ Max, or one of HUNDREDS of other businesses using Windows software discover, some too late, that their credit cards have been abused by hackers.

    Why do media sources continue to refuse to identify the OS compromised systems were running? Most depend on the ad revenues supplied by Microsoft and its partners to pay their salaries. They will not bite the hand that feeds them, regardless of the costs to others.

  4. SomeDude

    GaryGeek is exaggerating here. Yes, some extra effort is required to secure the most widespread operating system in existence. However, it is factually incorrect to say that every windows machine WILL become infected at some point. While you can lay some blame on Microsoft, the fact is that ALL computer software in existence has bugs. There is no perfect system and some responsibility does fall on the user, like it or not.

    Venturing off topic, Gary also mentions an old problem with the Win9X family. I certainly never knew anyone to claim an uptime of 50 days with older versions of Windows, but where might Gary have heard this? No doubt on a newsgroup or some forum dedicated to Linux where he spent countless hours arguing with Windows users. You can bet your life there was a lot of lying on both sides. In the digital sausage-measuring contest, the OS you are using is superior in every way.

    Whether or not you consider Linux is your choice, but beware of those who are so dedicated to one side that truth takes a back seat to advocacy. Gary’s mission is to spread Linux far and wide, and the very fact he doesn’t do it to earn a living should worry you a little.

Leave a Reply

Your email address will not be published. Required fields are marked *